Becoming a Grammar Teacher: Why? What? How?

By Richard Hudson
July 2012

Becoming a Grammar Teacher: Why? What? How?.....................................................1

Executive summary........................................................................................................1
1 Introduction and overview.....................................................................................6
2 My credentials........................................................................................................8
3 A brief history and geography of grammar teaching...........................................10
4 Why, what and how..............................................................................................13
5 Why?....................................................................................................................14
5.1 Writing.........................................................................................................15
5.2 Reading........................................................................................................16
5.3 Speaking and listening.................................................................................18
5.4 Foreign languages........................................................................................20
5.5 Thinking.......................................................................................................20
6 What?...................................................................................................................22
6.1 Lexical relations and morphology................................................................23
6.2 Syntax...........................................................................................................29
7 How?....................................................................................................................34
8 The obstacles........................................................................................................39

Executive summary
 This document is an attempt to make secondary English teachers more
enthusiastic about teaching grammar. Most such teachers received little or no
formal instruction in grammar either at school or at university, which is an
obvious and serious obstacle to the re-introduction of grammar teaching. But
negative stereotypes of grammar teaching discourage some teachers from
learning about it, so before this gap can be filled, it is possible that these
teachers need to be persuaded of the benefits of teaching grammar.
 Grammar ceased to be taught in our English lessons in about 1960, and
somewhat later in foreign language classes, but was reintroduced in the first
National Curriculum of 1990. Since then it has had an increasingly prominent
place in the curriculum for both English and Foreign Languages, including the
new (in 2012) draft curriculum for English.
 But as currently practised, grammar teaching does not seem to be working,
because (according to two grammar audits of new undergraduates) school
leavers don’t know any more grammatical terminology now than they did in
1986.
 This situation is neither natural nor inevitable, because at other times and in
other places grammar is, or has been, taught much more successfully.
 Why teach grammar? The research evidence and/or common sense suggests
that it benefits five educational areas:
o writing
o reading
o speaking and listening
 o foreign languages
o thinking skills
 What is there to teach in grammar? In the context of grammar teaching, it is
useful to distinguish two areas:
o lexical relations and morphology
o syntax
In both cases grammar can throw light on texts, whether literary or not, but
some kinds of analysis need a notation such as the diagrams that are often used
in linguistics.
 How should grammar be taught? Three general principles are offered:
o grammar should be taught systematically in a ‘spiral curriculum’,
which avoids the conflict between systematic teaching and teaching
‘when needed’.
o it should be integrated with other parts of teaching
o standard technical terminology should be used, as it is important for
different teachers, both within and across schools, to use the same
terminology.
Preface

Just after the National Curriculum was first introduced into England I was invited to
write a book about grammar teaching1. That was 1992, and twenty years later here I
am again, writing a book about grammar teaching in the National Curriculum. Why?
For one thing, the world has changed in the last twenty years. We’ve switched
governments from Tory to Labour, and then back to Tory again; and between them,
those governments have introduced three different versions of the National
Curriculum for English, and now we’re waiting for a completely new version of the
entire National Curriculum. Grammar was a significant part of the earlier versions,
but it’s likely to be an even more prominent part of the next one. As far as
Government is concerned, grammar is here to stay. But during those twenty years, a
lot has happened in our schools. The idea of having a National Curriculum has bedded
down and become widely accepted; the National Literacy Strategy has come and
gone; league-tables have come to dominate almost everything.
Other changes have affected the curriculum itself. The first English curriculum
is strong on variation, including differences between spoken and written language and
between Standard English and non-standard dialects2. But it gives very little detailed
guidance on what areas of grammar should be covered, or on what concepts and
terminology should be taught. The National Literacy Strategy introduced a lot of
grammatical guidance, including the first-ever official glossary of grammatical
terminology and a book of teaching suggestions about grammar for primary schools
called Grammar for Writing. The glossary was a remarkably poor document which
some of us linguists were allowed to revise long after it had been published;
fortunately, it seems to have been generally ignored but unfortunately, so does our
revision. In contrast, Grammar for Writing was widely welcomed and seems to have
had some effect on grammar teaching in primary schools. In contrast, there has been
very little official guidance on grammar for secondary teachers3.
One major change to secondary English teaching has been the rise and rise of
A-level English Language, which from tiny beginnings in the early 1980s has become
the 13th most popular A-level subject (and for girls, the 9th most popular)4. Thanks to
this new school subject which hardly existed in 1992, thousands of English teachers
are now teaching some grammar and presumably discovering that grammar might be
not only relevant, but also teachable and (even) fun.
A third change is that the UK is now recognised to be facing a crisis in its
foreign-language education. 1992, when my first book was published, happened to be
the year when A-level entries for foreign languages peaked, but since then the
numbers have declined dramatically, with French entries dropping from 31,000 to
13,0005. Similar falls have been registered in entries at GCSE so that fewer than half
of all pupils now take any foreign language at GCSE. Even more depressingly, these
problems only apply to state schools, so languages are in danger of becoming the
special preserve of independent schools. Almost everyone agrees that the crisis is

1
Hudson 1992
2
There is a list of relevant passages from the 1990 National Curriculum in Barton and Hudson
2002:263-79.
3
I was asked to produce two semi-official websites suitable for secondary teachers: ‘KS3 Grammar’
(http://www.phon.ucl.ac.uk/home/dick/tta/KS3.htm) and ‘Grammar in the Secondary National Strategy’
(http://www.phon.ucl.ac.uk/home/dick/kal/top.htm).
4
http://tinyurl.com/ca2010gce
5
For these and other figures, see http://www.phon.ucl.ac.uk/home/dick/ec/stats.htm#fl.
serious, to the extent that it has become an ongoing campaign issue at the British
Academy, and also that, among the many complex and deep-seated causes is the very
boring curriculum for GCSE languages. One possibly important element in this
gloomy picture is the very small role that grammar still tends to play in language
teaching. The idea of teaching a language without any mention of its grammar is a
bizarre one, with nothing to recommend it as part of education; and it is even odder if
children already know grammatical ideas and terminology when they come to the
Foreign-Language classroom, thanks to the National-Curriculum requirements for
English.
This brings us to a major contradiction in our schools. Ever since 1990, the
National Curriculum has required both primary and secondary schools to teach
grammar as part of English and Foreign Languages. But there are good reasons for
thinking that this bit of the National Curriculum has failed. Whatever teaching of
grammar may be taking place – and anecdotally one hears of a great deal of grammar
work being done in primary schools in particular – it doesn’t seem to be making much
difference. Teachers of A-level English Language, who are probably the best placed to
use any grammatical knowledge that pupils have picked up, report no increase at all6.
Similarly, a ‘grammar audit’ in 2010 of incoming undergraduates found that, if
anything, they knew less grammatical terminology than their counterparts did in a
similar audit conducted, using the same instrument, in 19867. In short, if you want
grammar to be taught, it’s not enough to say so in official documents; something else
is needed.
What is that extra something? Teachers, and especially secondary English
teachers, need to be both willing and able. The fact is that most English teachers were
taught no grammar either at school or during their undergraduate English degree8
(although the situation is very different now from 19929); and many came through
PGCE courses which actively discouraged them from teaching grammar. To say this is
not, of course, to criticise the teachers. The lack of grammar in their school and
undergraduate courses is not their fault, nor does ignorance of grammar imply lack of
ability. On the contrary, secondary English teachers are amongst the most able
teachers in our schools by any standard. Why should a teacher whose professional
expertise is based on literary and cultural studies take grammar (of all things!)
seriously, especially given the terrible press that grammar has had over the years?
This isn’t just a rhetorical question. Why should such a teacher bother to teach
grammar? Even more importantly, why should they bother to invest significant
amounts of time and pain in learning enough grammar to teach it confidently and
competently? This book is an attempt to take that question seriously, but I can already
hint at part of the answer: pupils like grammar. One rather trivial piece of evidence: if
I search the internet (in mid 2012) for the phrases “I hate grammar” and “I love
grammar”, Google finds twice as many examples with “love” than with “hate”. But a
much more significant piece is the enthusiasm among pupils for the recently

6
This is one of the recurrent themes of the major email list EngLang, which is dedicated to teachers of
English Language at A-level.
7
For the data and analysis, see http://www.phon.ucl.ac.uk/home/dick/ec/ba-kal/ba-kal.htm. I give a
little more information about this project in section 2.
8
Cajkler and Hislam 2002, Williamson and Hardman 1995
9
But PGCE courses are much more willing nowadays than they were in 1992 to accept graduates of
either English Language or Linguistics. For some research data on requirements for English PGCE
courses, see Blake and Shortis 2010, Committee for Linguistics in Education and others 2010; an
earlier survey is discussed, with data, at http://www.phon.ucl.ac.uk/home/dick/ec/pgce-clie.htm.
introduced Linguistics Olympiad10, which is largely about grammar. As one pupil
said, "it mashes yer 'ed, but in a good way".

10
See http://www.uklo.org.
1 Introduction and overview
This book is a plea for schools to teach grammatical analysis as a standard and
significant part of the curriculum. To give a taste of what I have in mind, let’s move
straight to the end of the story, to a school where grammatical analysis is already well
established and working smoothly. It could have been a primary school, but a
secondary school is more relevant to this book.

*****

Let’s follow a Year 8 class in this school. The first lesson is English, which
starts with a brief activity about modal verbs: What’s the difference between must and
have to, as in I must go and I have to go? The students work in pairs to spot the
differences, and after five minutes the teacher compiles a list of contrasts such as to
on have, s on has in He has to go, and the possibility of Having to go but not of
Musting go. This leads into a discussion of how to choose between the two when
writing, which then broadens out to include other modal verbs such as may and can,
and the differences between them. The discussion is a discussion among equals
because all the students use these verbs every day, so they are already experts; what
they like is explicit awareness of the choices they have to make.
The next lesson is French, which again starts with a warm-up exercise. This
time it’s about how possession is expressed. The teacher presents half a dozen French
sentences such as Jean touche la valise de Marie and Jean touche la main à Marie.
The question for the students is how to choose between de and à. Once they work out
the difference between ‘alienable’ and ‘inalienable’ possession, the teacher directs
their attention to English sentences like John touched Mary on the hand, and asks if
this might have something to do with the same contrast. By the end of the lesson, the
students have a good understanding not only of terms such as ‘possession’ and
‘alienable’, but also a better understanding of how languages can be both similar and
different at the same time. They have also, on the way, learned and practised the
French patterns, but they will return to these in the next lesson.
Then they move into a maths lesson. The subject is algebra, and in particular
the technique of expanding brackets as in 3 x 63x 18. The teacher, who (of
course) is well-versed in grammar, points out that something very similar happens in
coordination, where He sang and danced can be expanded into He sang and he
danced. He reminds them of the notation for coordination which uses brackets, as in
He (sang and danced). This incident only occupies a few minutes in the middle of the
lesson, and illustrates the benefits of being able to call on grammatical understanding
at any point.
Similarly, history mentions some grammatical points in an old document, and
chemistry mentions the similarity (familiar to linguists) between the ‘valency
bonding’ of atoms and the ways in which words hang together in a sentence. He also
shows how the word “polyvalent” can be analysed into “poly”, “val” and “ent”.
Citizenship mentions the grammatical choice between first and second person as the
basis for any face-to-face interaction, and biology touches on the debate about
whether grammar is innate or learned.
Several teachers comment on grammatical issues in children’s written work.
The history teacher briefly mentions the choice between “Henry the Eighth’s second
wife” and “the second wife of Henry the Eighth”, and the chemistry teacher points out
how the link between “valent” and “value” helps to remember that both words have
just one ‘l’.
Grammar, in short, pervades the whole curriculum in much the same way as
numeracy and literacy.
But that’s not all, because enthusiasts go off at lunchtime for the weekly
‘Linguistics Club’ run by the French teacher, who is preparing for the annual UK
Linguistics Olympiad11. This is an academic competition for schoolchildren like the
Maths Challenge, but focusing on language problems. The challenge is to work out
how the language concerned ‘works’, and some (but not all) of the problems involve
grammar. This week the problem is from Ulwa, a Nicaraguan language, where the
earlier discussion of French possessive patterns prepares them to work out the much
more complex system of Ulwa, where siknibilh, ‘our (inclusive) horsefly’ consists of
sikbilh with ni inserted right in the middle. Among other things, students learn to talk
comfortably about infixes.

****

At this stage in the argument I would expect some readers to roll their eyes
and wonder what planet I’m on. Surely it’s ridiculous to imagine so many people
knowing so much grammar? And especially so in the harsh reality of 2012, where
most teachers have themselves come through a completely grammar-free education at
both school and university?
In any case, is this a good use of resources, compared, say, with basic literacy
or numeracy? What’s the point of training the population to think about grammar
when the bills aren’t being paid?
I recognise these objections and take them seriously, but the point of this book
is to confront them and answer them. The first objection can be dealt with relatively
easily. The scenario I described is totally realistic, in a global sense, because it’s a
minor variation on what happens already in some countries. For instance, a recent
description12 of grammar teaching in The Netherlands reported that ‘teachers seem to
like to teach grammar’, so they do more of it than is strictly required by the
curriculum. For such teachers, it is very easy to imagine them including warm-up
activities on grammar, and freely using grammatical terminology in the same way as
they might use terminology from any other subject area. Nor is the ‘Linguistics Club’
fantasy – it’s already happening in the UK. Some schools really do run just such a
club, and do enter children for the UK Linguistics Olympiad. In fact, as of 2012, two
thousand children and 300 schools are involved in the Linguistics Olympiad at some
level, and the numbers are rising fast.
The second objection is harder to counter: even if grammatical analysis on this
scale is possible, is it desirable? And in particular does it justify the considerable cost
of bringing teachers up to speed in grammar? These are the main questions for this
book.

11
http://www.uklo.org.
12
Position paper on ‘The place of grammar in the L1 curriculum in the Netherlands’,
by Gert Rijlaarsdam, 2011; prepared for ‘Stakeholder Conference: Re-thinking
grammar and writing’, 3rd February 2012
 It should be clear from this scenario roughly what I mean by ‘grammatical
analysis’, but just to avoid misunderstanding I’ll now spell out what I am not
advocating.
I’m advocating grammatical analysis, which is an important part of linguistics,
but that’s as far as I go in this book: I am not advocating a general course in
linguistics. Of course, as a linguistician13, I would love to see every child covering a
lot of elementary linguistics – how children succeed in learning language but
chimpanzees fail, how dialects evolve and differ, how we make consonants and
vowels with our mouths, how vocabulary divides up the world in a systematic way. I
also believe passionately that education would be greatly improved by a stiff dose of
elementary linguistics. But that’s not the argument I’m making here. This book really
is just about grammar. I’ll explain more precisely what that means in section 6, but for
the present it’s enough to say that it’s about how words are organised to convey
complex meanings.
Nor am I advocating a return to prescriptive grammar. Prescriptive grammar
‘prescribes’ (recommends) some forms and ‘proscribes’ (bans) others; it prescribes I
haven’t and proscribes I ain’t; it prescribes the book in which I found it and proscribes
the book I found it in; and so on. In contrast, grammatical analysis accepts any form
that’s in regular use, and analyses it. That doesn’t mean simply stamping it as ‘ok’; it
means pursuing all its characteristics, including the social differences between haven’t
and ain’t, and trying to understand the dynamics of the system – how ain’t fits into the
larger system of auxiliary verbs. I reject prescriptive grammar totally, as really bad
science; but I don’t reject the goal of teaching standard English as a supplement to
local non-standard varieties. I just think that prescriptive grammar isn’t just bad
science, but it’s also bad pedagogy because it rejects outright so much of what
children already know. Instead of undermining this knowledge, we should be building
on it as a firm foundation. I have more to say about the flaws in prescriptive grammar
in section 3.
Nor am I advocating a simple return to the grammatical analysis that used to
exist in some schools, and that still happens in some. Grammatical analysis was part
of the official curriculum till some time in the 1960s, but since it died in most schools,
grammatical analysis has been re-born in our universities and grown spectacularly.
There’s no need to look backwards for a model, because linguistics offers many
excellent models (section 6).

2 My credentials
This is a very personal statement of beliefs, so readers need to know how seriously to
take it. At this point I face a dilemma, because English people are deeply reluctant to
boast. How does one present one’s credentials without boasting? I take comfort from
the fact that my first ‘boast’ is that I’m a grammarian, which for some people isn’t a
boast at all, but a confession.
My credentials in this campaign start with my years at secondary school,
where I learned a lot of grammar and enjoyed it immensely. Those years, and that
expert teaching, turned me into a grammarian for life, so after a remarkably grammar-
13
My colleagues in linguistics like to call themselves ‘linguists’, but they also grumble about the
confusion that this term causes by lumping them together with the ‘linguists’ of the worlds of education
(where anyone who studies a language is a linguist) and business (where the Chartered Institute of
Linguists exists for professional translators, interpreters and others who make a living from language).
The solution lies in our hands: if we call ourselves ‘linguisticians’ nobody will ever confuse us with
linguists of other kinds; and, as a bonus, it will emphasise our similarity to phoneticians.
free undergraduate course in French and German, I did a PhD on grammar (in this
case the grammar of an unwritten language spoken in the Sudan, Beja) and then spent
the rest of my working life teaching and researching grammar at UCL14. So I know a
great deal about the world of grammar as practised at university by linguisticians – a
world which has had hardly any impact on school teaching.
Although a linguistics department can be very much like an ivory tower, I
actually spent a lot of time looking out at the very different world of schools. This was
largely because I worked for a few years with a brilliant bridge-builder, Michael
Halliday, whose work with schools I very much admire15. Following his inspiration, I
was a founding member in 1980 of the Committee for Linguistics in Education16,
whose mission was (and still is) to build bridges between linguistics and schools; and
as of 2012 I’m an active member of this committee (which celebrated its hundredth
meeting in early 2012). But the work I take most pride in is the UK Linguistics
Olympiad (which I mentioned earlier), where I chair the committee that CLIE created
in 2009.
I’ve also worked quite a lot with government departments on documents such
as curriculums and materials about grammar. I had a hand in a heavily revised
glossary for the National Literacy Strategy17, in a very well-received book for primary
teachers called “Grammar for Writing”18, in guidelines for teaching foreign
languages19, in the specifications for the ill-fated Diploma in Languages20, and in the
grammar component of the new National Curriculum for English21 promised for 2013.
I have also written training material about grammar for secondary teachers22, and a
rationale for the grammar component of the old Literacy Strategy23. So I know what
the existing government documents say about grammar.
Unfortunately I also know that there’s a gulf between government documents
and the reality in the classroom. In 1986, a colleague and I tested the very elementary
grammatical knowledge of the school-leavers who were most likely to know any
grammar, namely the new undergraduates in a department of linguistics or of foreign
languages. The test was very simple, consisting of a sentence in which students had to
find a noun, a pronoun, a conjunction and so on through fifteen word classes. This
little project took place before the 1990 National Curriculum first laid down the
minimum standards for grammatical knowledge, so when I repeated it (with a
different colleague) in 2010 we should have found a significant increase in
knowledge. In fact, we found that students knew even less grammatical terminology
in 2010 than they had in 1986. I mention this little project as evidence that I’m
enough of a realist to recognise the formidable obstacles that have to be overcome
between where we are now and the imaginary world of section 1. I discuss some of
these obstacles in section 8.
One important qualification for writing this book is breadth of both vision and
experience. As a linguist, I see the whole of language education as a single complex

14
My website is http://www.phon.ucl.ac.uk/home/dick/home.htm.
15
Halliday 2007
16
http://www.clie.org.uk
17
This revised glossary is included in one on my website:
http://www.phon.ucl.ac.uk/home/dick/tta/glossary/KS2&3.htm
18
Anon 2000
19
The guidelines are no longer available online.
20
Launched in 2008 and abandoned in 2010; see http://www.linksintolanguages.ac.uk/news/786.
21
http://www.education.gov.uk/schools/teachingandlearning/curriculum/nationalcurriculum.
22
http://www.phon.ucl.ac.uk/home/dick/tta/KS3.htm
23
http://www.phon.ucl.ac.uk/home/dick/kal/top.htm
network, whether it concerns a child’s first language or a foreign language. (Among
linguists and educators this view is called ‘Language awareness’24.) Remarkably few
government documents bring these two strands of teaching together (unlike my own
schooling, where the same teacher taught English and Latin). Indeed, even within
first-language English teaching there seems to be a significant division, at least in
research, between thinking about reading, writing, speaking and listening; this
division is crucial when discussing grammar because grammar is much more closely
associated, in teachers’ minds, with the teaching of writing than of any other modality.
As a bridge-builder, I also see universities as part of the same educational system
as schools, with universities receiving students from schools but also providing
teachers and expert knowledge. Again this holistic view is rather rare, especially in
universities (which tend to see themselves as victims of the school system rather than
as colleagues).
One final qualification for writing this book takes me back to the research world
of linguistics, and of grammar in particular. This world has been growing and
changing very fast since the 1950s, with two kinds of product to show for all the
activity. On the one hand, we now know a great deal about grammar, and in particular
about the grammar of English, which we didn’t know sixty years ago. And on the
other, we have a large collection of general theories of grammar – how sentences and
words are structured, and how grammars can say what’s possible and what isn’t. I
know quite a lot about these theories, so I know which ideas are safely
uncontroversial and where the theoretical minefields lie. (In fact, in 1980 I compiled a
list of 83 claims that were uncontroversial in linguistics25.)
My view is that the last thing schools need from linguistics is to get mired in our
research controversies, but I also have to declare a vested interest because I myself
invented one of these packages (called Word Grammar). Not surprisingly, I believe
that Word Grammar is better than any of the others; but I also believe that it has
something rather special to offer schools which no other theory does. I’ll explain this
in section 6, where I shall also note that my suggestion is controversial among my
professional colleagues.
I would like to think, therefore, that as an elderly, experienced, research-expert
and broad-minded academic I have some hope of offering the ‘joined-up thinking’
that seems so hard to achieve, and that is certainly quite unusual in the area where it is
arguably most needed: education.

3 A brief history and geography of grammar teaching

Grammar has an extremely respectable ancestry. The earliest evidence we have26 is
from nearly 4,000 years ago, when the Akkadian-speaking scribes of Babylon learned
to translate into Sumerian (which by then was already dead). Their training included
learning tables of equivalent verb-forms in the two languages, so someone must have
analysed these verb forms and produced a systematic framework. Rather remarkably,
they ordered first, second and third person forms in that order, so that particular part
of our heritage may be four thousand years old – a spectacular example of scholarly
transmission.
The term ‘grammar’ comes from the Greek expression grammatike tekhne,
meaning "art of letters," which also contains gramma "letter", so its modern meaning
24
The best place for an introduction to Language Awareness is
http://www.languageawareness.org/web.ala/web/tout.php.
25
Hudson 1981
26
Gragg 1994
is a narrowing of the original, though it is still closely associated with writing. The
Greeks developed the tradition of grammatical analysis that dated back to the
Babylonians into a more highly structured and theoretical system – or, more
accurately, a series of different and competing systems – which linked not only to
school teaching but also to philosophy27. Somewhat later, the Romans adopted this
legacy and applied it to Latin, forming the basis of the European grammatical
tradition which survived, with remarkably little change, into the nineteenth century.
For the Greeks and Romans, the school curriculum (called ‘the liberal arts’)
had just three parts, one of which was grammar. (The other two were logic and
rhetoric.) This tradition persisted through the Middle Ages, with Latin still as the
medium of instruction; so grammar was essentially the grammar of Latin, rather than
of English. Grammar dominated the entire curriculum, a fact which we celebrate in
the name we still give to some of the schools which were founded in the late Middle
Ages (or their more recent equivalents): ‘grammar school’. By the nineteenth century
the school curriculum had broadened considerably, but in public schools and grammar
schools grammar still played a significant part in the teaching of foreign languages
(modern as well as classical) and in the teaching of English.
How significant is ‘significant’? In foreign-language teaching the dominant
approach is now called the ‘grammar-translation’ method28, a name which reflects the
centrality of grammar. Another indication of significance is the supply of textbooks,
which in the early twentieth century were solid, serious, rather dull and widely used.
For example, in French the dominant textbook author from at least the 1930s to the
1960s was W. F. H. Whitmarsh, M.A. (the ‘M.A.’ is important as a sign of academic
reliability), who wrote several dozen textbooks for schools. English grammar was
similarly dominated by a single figure, J. C. Nesfield, M.A. (again), whose ‘Manual
of English Grammar and Composition’29 was first published in 1898 and was so
influential that it spawned a further generation of derivative textbooks. Nesfield’s
textbook is nothing if not solid: 418 pages of tiny print, with the first hundred devoted
to grammatical analysis. The detailed discussion of word classes has a clear and
ambitious purpose: to prepare students to apply a standard grammatical analysis to
any sentence a teacher or examiner might throw at them. The only other subject that
had, or has, a similarly ambitious goal is mathematics. One interesting characteristic
of Nesfield’s grammar is the standard table for laying out the parts of a sentence; this
is an example of the notation for which I shall argue in section 6.
This educational tradition was shared, by and large, by all of Europe, and
indeed it was exported to the overseas colonies and territories. In many of these
countries, grammar still has its traditional status and content, albeit with some features
modernised. In most of Eastern Europe, school children still spend significant
amounts of time learning how to analyse words and sentences in their own language,
and build on this grammatical understanding when learning foreign languages; and
the same is true in most countries whose language is descended from Latin (such as
Italy, Spain, Portugal and France, and their overseas extensions)30. It’s hard to know
how successful this teaching is, especially without knowing what its precise goals are;
but one measure is popularity among adults who ‘did some grammar’ at school. The

27
Robins 1967
28
For a brief discussion and exemplification, see http://www.nthuleen.com/papers/720report.html.
29
Nesfield’s book can be read online at http://archive.org/details/manualofenglishg00nesfuoft.
30
For instance, a group of undergraduates from a Spanish university took a test (in English) that we
used for assessing UK undergraduates’ knowledge of grammatical terminology, and outperformed all
the UK groups.
fact is that a lot of people hated it, but (by an admittedly crude measure31) even more
loved it.
However, the UK was different. In this country, grammar-teaching more or
less disappeared in the 1960s, as it did in most other English-speaking countries32. The
reasons for ‘the death of grammar’ are complex and deserve more research, but one
element in the explanation is certainly the lack of grammatical research in our
universities throughout the early twentieth century33. This left universities with
nothing to teach their undergraduates about grammar, and therefore no intellectual
boost for future school teachers comparable to the updating and rethinking that
undergraduates receive in other subjects.
The result was a decline in the teaching of grammar in English lessons, with
teachers applying half-remembered analyses from their own school days and using
textbooks based directly on the previous generation of school textbooks, without any
academic input. Some teachers still inspired (as mine did), and some children still
enjoyed their grammar classes (as I did); but most grammar lessons were boring,
dogmatic and intellectually frustrating. It is hardly surprising that English teachers
started to ask what the point was, and welcomed with open arms a series of research
projects which showed that grammar lessons had no impact at all on the quality of
children’s writing34. Since the standard argument for grammar was that it improved
writing, this research was the end of grammar teaching – much to the relief of a great
many people who had been campaigning against grammar and in favour of literature.
The effect was that in the early 1960s the one remaining optional question about
grammar was removed from the O-level English paper, and, effectively, grammar died
in English. Seen from a global perspective, English (and England) had entered the
Grammatical Dark Ages. Meanwhile, of course, all was by no means doom and gloom
in English teaching, “in which at last a subject in the curriculum
engaged with students’ voices, imaginations and responses in a way
which is widely seen as being one of the most positive
developments in school subject teaching of the 20th century.”35
Meanwhile, of course, grammatical knowledge was needed in foreign-
language teaching so long as this was dominated by the grammar-translation
approach. This kind of teaching became increasingly difficult as more and more
English teachers abandoned grammar, but foreign-language teaching had its own
agenda, and grammar-translation gave way to other approaches in which grammar
was less central. The new ‘communicative’ syllabus, with its focus on knowing how
to carry out very specific tasks in the target language, allowed teachers (and text-book
writers) to replace grammar by memorized phrases. Although more recent research
has confirmed that students learn foreign languages better if teaching focuses
explicitly on grammatical or lexical forms, with or without attention to meaning,36
foreign languages are often taught without any use of grammatical terminology. It is
interesting to wonder whether foreign-language teachers would have adopted the
31
In May 2012, Google found 77,000 examples of ‘I hate grammar’ compared with 162,000 of ‘I love
grammar’.
32
Australia and New Zealand had a similar history to the UK. Although the USA shared in the grammar
debates of the 1960s, grammar teaching seems much more common in English lessons there than here;
but it also seems to be more focused on avoiding ‘errors’ than on grammatical analysis per se. I don’t
know the state of grammar teaching in Canada, Ireland or South Africa.
33
Hudson and Walmsley 2005
34
Andrews and others 2004b, Andrews and others 2004a, Andrews 2005
35
Gary Snapper, personal communication
36
Norris and Ortega 2000
communicative approach so enthusiastically if their pupils had come to them with a
good supply of grammatical terminology learned in English.
The main point of this brief history and geography is that normal schooling
does include grammar teaching, so there is nothing ‘normal’ or ‘natural’ about
ignoring it. On the contrary, grammar played a central role in Western education for
thousands of years, and continues to do so in many countries. We are the ones who are
out of step, and the term ‘Dark Ages’ that I used earlier really is justified as a
description of the grammar-free education that most of our children have received.
However, this situation can be seen as an opportunity for a new start.
Paradoxically, while grammar has been ignored in schools, it has flourished in
university research and teaching. In 1921 it was possible to write37 that it was “…
impossible at the present juncture to teach English grammar in the schools for the
simple reason that no-one knows exactly what it is…”; nearly a hundred years later,
we know a great deal about English grammar, thanks to a series of block-buster
research-based grammars.38 At the same time, a good deal of linguistic thinking (short
of grammatical technicalities) have found a place in schools, so teachers are used to
teaching about non-literary genres, spoken language and variation. Most importantly
of all, perhaps, Standard English has taken its place among a range of alternatives
which are accepted as equally ‘correct’ in their own terms. With this background, both
academic and pedagogical, a new version of school grammar can be developed which
is much better than what was taught in the nineteenth century.

4 Why, what and how

I now turn to the case for grammatical analysis – what I shall now call simply
‘grammar’, though I shall spell out the exact meaning in section 6. Why should it be
taught, what should it consist of, and how should it be taught? These three questions
are obviously closely connected.
For example, if the main reason for teaching grammar is to allow English
teachers to deal with ‘common grammatical errors’, then we immediately have an
answer to ‘what?’: grammar consists of whatever terminology is needed in order to
define these mistakes. For instance, if one of the errors is so-called ‘split infinitives’,
then children should be taught to recognise infinitives and name them. This in turn
suggests an answer to ‘how?’: rather didactic lessons in which the teacher goes
through errors one at a time, providing an example, a name for the error, a definition
of the error and its correct alternative, possibly an explanation for why it is an error,
and then an exercise in which students spot examples of the error and correct them.
Needless to say, this is not the kind of teaching that I have in mind, but that is as
much because I reject the initial motivation (why?) as because I reject the content
(what?) and the method (how?).
What I shall argue below is that there are many good answers to all three
questions, though the answers are to some extent interconnected in the way suggested
above. It will be helpful to separate out the different answers to each question, if only
for intellectual clarity, though in some cases it may be possible to combine answers;
for example, it is surely possible to combine different aims in a single lesson. But it
will certainly not be helpful to continually look across from one question to the
others. In discussing the educational benefits of grammar (why?) it will only be
confusing to bring in further questions about the content of the grammar (what?) and
37
Board of Education 1921
38
Quirk and others 1972, Quirk and others 1985, Biber and others 1999, Carter and McCarthy 2006,
Huddleston and Pullum 2002.
the pedagogical methods (how?). These links are important, of course, but for my
purposes they will complicate the argument unnecessarily so, by and large, I shall
ignore them.

5 Why?
Why, then, should grammar be taught? For some teachers, the only driving force is
‘because The Boss says so’, where The Boss could stand for various figures ranging
from Head of Department, through Head Teacher, to the Secretary of State for
Education. This is to be expected, given the powerful campaign that finally removed
grammar from the curriculum. For decades, teacher-trainers have been telling trainee
teachers that teaching grammar is a waste of time, or worse, and it is common to hear
grammatical analysis described as merely ‘the naming of parts’, a pointless exercise
in classification. As I shall explain in section 6, such comments miss the point almost
entirely, but the fact is that these views are influential.
There are, in fact, some very good reasons for teaching grammar, and
improving writing is only one of them – an important reason, but by no means the
most important one. The debate about the pros and cons of grammar teaching had
such a negative outcome in part because it focused almost exclusively on this reason,
ignoring all the others – improved reading, foreign-language learning and general
thinking skills. Imagine a similar debate about the pros and cons of teaching algebra
which focused entirely on the benefits for domestic account-keeping and concluded
against algebra on the grounds that it doesn’t help people to avoid debt. One of the
purposes of this book is to restore the balance in the debate about grammar by
presenting a broader picture.
One important consequence of looking beyond any one aim of grammar
teaching is to reframe the debate about the benefits of teaching grammar. Like any
other teaching activity, teaching grammar has a cost, in terms of time that could be
devoted to other worthwhile activities, in terms of teacher training and in terms of
pupil commitment. The debate is about whether this cost is balanced, at least
subjectively in the eyes of teachers, by educational benefits to the pupils, so it is
important to have a clear global view of both the costs and the benefits. If grammar
teaching produces benefits in just one area, such as writing skills, these benefits must
balance the total cost of teaching grammar from scratch. But if the benefits are spread
across a wide range of areas, the cost per area is clearly much lower.
To see the importance of this point, consider two contrasting scenarios. In the
actual current scene, children know very little grammar when they leave primary
school, so any secondary teacher who wants to mention grammatical concepts must
teach not only the concepts concerned, but also a whole raft of elementary ideas
which underpin them. Historically, this is the situation in which foreign-language
teachers found themselves in during the 1970s when English teachers stopped
teaching grammar, and it is hardly surprising that grammar-free teaching methods for
teaching foreign languages became so popular. Now consider the scenario that exists
in some countries, in which children learn and consolidate a great deal of grammar in
primary school. By the time they reach secondary school, basic grammatical concepts
and terminology are familiar, so there is no cost at all to using them – indeed, the
more they are used, the more firmly they become entrenched in children’s minds.
Admittedly a secondary teacher may need a specialised concept that the children have
not yet learned – such as ‘modal verb’, ‘subjunctive’ or ‘ablative’ – it will be
relatively easy to teach because children already understand foundational concepts
such as ‘verb’, ‘subject’ and ‘subordinate’. In this scenario, grammar is virtually cost-
free in secondary schools, and its costs in primary are offset against a wide range of
benefits in secondary.

5.1 Writing
The first reason, then, is the one that most educationalists have concentrated on for the
last few decades: teaching grammar improves first-language writing skills. The
argument is that mature academic writing (the target of school literacy teaching)
requires high-level linguistic skills, including not only a broad vocabulary but also
sophisticated grammatical skills. These skills are of two kinds, negative and positive:
 standardness, meaning the avoidance of forms from the local Non-standard
dialect (e.g. ain’t); this is sometimes called ‘accuracy’ or ‘correctness’.
 diversity, i.e. the sensitive use of a wide range of constructions, including
constructions that aren’t normally used at all in ordinary conversation (e.g.
While working in the garden he injured himself).
Until very recently this argument has carried very little weight in the English-
speaking world because of the research (mentioned earlier) that purported to show
that teaching grammar simply did not work as a way of teaching either kind of skill.
However this research had a fundamental flaw: all it showed was that grammar can be
taught ineffectively. Typically, a class would have (say) a weekly lesson on grammar,
and their written work would be compared with that of another class that had no such
lesson. The results showed that grammar is ineffective when taught in this way; but it
did not show that this was the only possible way to teach grammar.
The received wisdom has been overturned by two recent strands of research,
both conducted in Britain. Since the 1990s, the psychologists Peter Bryant and
Terezinha Nunes and their colleagues in Oxford have shown that explicit instruction
in morphology (the grammar of word-structure) does indeed produce measurable
positive effects on children’s spelling, their use of apostrophes, and the growth of
their vocabulary39. For example, children were better able to distinguish plurals and
possessives in pairs such as boys and boy’s after practising morphological analysis
than when the practice involved just pronunciation or just meaning and syntax
(sentence structure)40. More recently, the educationalist Debra Myhill and colleagues
in Exeter have shown considerable benefits in a large-scale study from ‘focused’
teaching of specific grammatical patterns; for instance, discussion of modal verbs
such as may and must produced benefits in the children’s use of modal verbs in their
own writing41. Teaching is focused, in this sense, if it concentrates on patterns which
are then tested in the children’s writing. This is an important qualification, which I
return to in section 7, because one of the main problems with previous research was
that the teaching was unfocused, so what was tested bore only a very general relation
to what was taught – a rather obvious weakness.
Another important strand of research, this time from America, gives somewhat
weaker support for teaching grammar42. This research shows that a classroom activity
called ‘sentence combining’ is good for the children’s writing skills. In sentence
combining, the teacher provides two or three single-clause sentences for the class to
39
Nunes and Bryant 2006
40
Bryant and others 2004
41
Myhill and others 2010
42
Andrews and others 2004b
combine into a single sentence. For instance, given the sentences (1) The boys were
playing with the dog. (2) They were standing on the pavement. (3) It was barking
loudly, a class might synthesize a wide range of sentences including the following:
 The boys who were standing on the pavement were playing with the dog that
was barking loudly.
 The boys standing on the pavement were playing with the loudly barking dog.
 While standing on the pavement, the boys were playing with the dog barking
loudly.
 Although it was barking loudly, the boys standing on the pavement were
playing with the dog.
The research shows that this activity has a strong positive effect on writing quality.
The only uncertainty is about whether the activity can really be called ‘teaching
grammar’. Most of the research literature rightly contrasts it with grammar teaching
as this is normally interpreted, at least in the States. But as I shall explain in section 7,
there are many ways of teaching grammar, and a good case can be made for
recognising sentence combining as one of these methods, whether or not the teacher
or students use technical grammatical metalanguage.
I introduced this discussion of writing skills by distinguishing negative
standardness and positive diversity. Both are legitimate targets of grammar teaching if
the aim is for every school leaver to be able to write mature standard English, but the
research that shows the positive effects of grammar teaching has focused on diversity
rather than standardness. This is reasonable because the two goals are very different.
Teaching children to write isn’t rather than ain’t is intellectually very easy, but may
raise emotional problems for children who use ain’t in their family, especially if ain’t
is labelled simply ‘wrong’ (or even worse, ‘bad’). This kind of teaching is often based
on a list of ‘common grammatical errors’ which has been handed down from one
generation of school teachers, via their pupils, to the next. In contrast, teaching
children to use a wider range of modal verbs is intellectually difficult (because of the
subtle meanings involved), but emotionally easy since it doesn’t threaten the
children’s identity. Moreover, it is rather obvious that telling children to avoid ain’t in
writing does have an effect, because most school leavers do avoid it in writing even if
they use it in speech; whereas the received wisdom was that merely telling children
about modal verbs would have little or no effect on their use of modal verbs. The
main point to emerge from this subsection is that this is wrong. Focused teaching of
specific grammatical points does indeed increase the diversity of children’s writing.

5.2 Reading
Another argument for teaching grammar is in improving reading skills. This is where
it is important to stress that ‘grammar’ means grammatical analysis rather than mere
error-avoidance. If the teaching of grammar was all about avoiding forms such as
ain’t, it would be irrelevant to reading; after all, it is the author rather than the reader
that chooses what words to use. In contrast, grammatical analysis is highly relevant to
reading because it is simply a conscious and articulated version of the analysis that
any reader makes. To read a sentence is to analyse it – its words, its grammar and,
ultimately, its meaning. The only differences between ordinary expert reading and
grammatical analysis are that the latter is completely conscious and reflective, that it
is explicitly expressed in terms of a standard terminology and notation, and (of
course) that it is much, much slower.
Grammatical analysis helps children’s reading in two ways, one very specific
and the other global. Specific grammatical instruction helps children when reading
sentences with particularly difficult syntax. The evidence for this claim comes from
research by Ngoni Chipere43, working with non-academic 18-year olds. The task for
his subjects was to read complex sentences like The doctor knows that the fact that
taking good care of himself is essential surprises Tom, and then to answer questions
such as What does the doctor know? or What surprises Tom? All the sentences had a
similar syntactic structure, so it was possible to train some subjects in handling
sentences with this structure by showing them how to break the sentences down into
simpler sentences and then to recombine these. The experiments showed that this
training had a clear positive effect on comprehension skills, so trained subjects
understood the sentences better than untrained subjects did. Like the research on
writing reported above, this research shows the benefits of highly focused grammar
teaching in which the construction taught is also the one tested in the research.
Turning to the global benefit of grammar teaching for reading, this is more a
matter of conjecture than of research, but it is so plausible that it should be taken
seriously. The claim is that grammar teaching encourages pupils to pay more attention
to grammatical structure in their reading – in other words, to ‘notice’ it44. When we
meet a difficult or unfamiliar word or structure, we can react in different ways. At one
extreme, we can stop and look at it to puzzle out what it means and how it works. This
is more likely if we are used to grammatical analysis because that is precisely what
such analysis consists of: working out how the parts of a word or sentence fit together
and produce their joint effect. At the other extreme, we can give up on grammar and
guess the meaning. This is what most of us do with sentences such as this: No head-
wound is too trivial to be ignored45. This apparently simple sentence is remarkably
hard to understand, as you will confirm by considering its opposite: No head-wound
is too trivial to be treated. I would guess that you found this just as sensible as the
first, but they can’t both be sensible. In such cases, we switch off the normal rules of
grammar and rely on common-sense.
Between these two extremes are constructions that most of us read but don’t
hear, such as the house in which he lives. If the goal is simply to extract meaning, a
reader can simply ignore the position of in; but a reader who finds syntax interesing
will pause and absorb the syntactic detail. There is ample evidence that syntactic
knowledge grows throughout the school years46, and that one of the main models –
perhaps the main model – for young writers is the material that they read. Moreover it
is clear from errors such as the house in which he lives in that novices struggle to
understand sophisticated constructions; so the more attention they pay to the
sentences they read, the more effectively they will learn to use such constructions
themselves. But paying attention to syntax is a waste of time if the only aim is to
extract meaning. Some children are natural ‘noticers’, but many are not; and (so the
argument goes) it is for the second kind of child that grammar teaching is particularly
important. Of course the grammar teaching for writing may focus on particular
constructions, such as in which, which children need to learn, but there is simply too
much grammatical detail to teach it all in classroom time, so the best strategy is to
provide a general tool which will allow children to learn from their reading.
One particularly important kind of reading where grammatical analysis is
especially helpful is the reading of literature – stories, novels, poems and so on.
Literary works that are read in class are, by definition, well written, so they serve as

43
Chipere 2003
44
Keith 1999
45
Wason and Reich 1979
46
Perera 1984, Hudson 2009
an excellent model for linguistic novices. And indeed, one of the main arguments for
linking literature to language-teaching in English is that children will become better
writers through reading literature. As we all know, mere ‘exposure’ works for some
children, but not for all, and maybe those for whom it does work are those who are
naturally inclined to ‘notice’ the grammar and vocabulary of what they read. If so,
then small amounts of carefully focused grammatical analysis may help the others to
notice, and learn. I give an example in section 6 of this kind of analysis.

5.3 Speaking and listening

Although grammar is historically associated with the written language – after all, in
Greek gramma meant ‘letter’ – it is highly relevant to the spoken language as well
because this is the source of written language. Spoken language, including the most
spontaneous and casual conversational styles, is controlled by much the same
grammatical rules as the most formal writing, even if real-time production allows
slips of the tongue and disfluencies that would be edited out in writing. A sentence
such as I love you comes out the same whether written or spoken, and has different
syntax from its equivalent in other languages such as French Je t’aime, Latin Te amo
or Arabic Ahibbik.
One of the very positive recent changes in our school curriculum is the much
higher profile that we now give to the spoken language in both first-language English
and foreign languages. Grammatical analysis has an important contribution to make in
teaching about spoken language. I consider here first-language English, leaving
foreign languages till the next subsection, and distinguish three different kinds of
contribution: transcription, status-raising and detail.
Transcription of spoken language is an important exercise for any child47. It is
easy to record conversation, and given a recording, any child can transcribe it; and of
course if that child happens to be a participant in the conversation, so much the more
interesting for them. One of the lessons that emerges from this activity is that any
transcription goes well beyond the purely phonetic substance of the recording. The
transcriber makes decisions about words (e.g. wait or weight?) and about grammatical
structures (e.g. fun and games or fun in games?), so a transcription is, in effect, a
grammatical analysis. This is particularly true if the transcription includes
punctuation, since this forces decisions about sentence-boundaries and other major
structural distinctions. Discussing these decisions in class is a useful way of
sharpening pupils’ awareness of grammatical structure, but (of course) it presupposes
a shared framework of ideas and terminology – precisely what grammar teaching
provides.
Status-raising is particularly important for the large majority of children –
probably about 80-90%48 – who natively speak a non-standard variety of English.
Thanks to the very negative attitudes to non-standard forms (described variously as
‘wrong’, ‘careless’ or ‘bad’), schools traditionally left non-standard speakers with
quite unnecessarily negative feelings about how they spoke, described by some as
‘linguistic self-hatred’49. Even if schools aim to teach everyone to speak Standard
English when needed, there is no reason why this should not co-exist with the local

47
For guidance, see the website produced for BT by Julie Blake and Tim Shortis at
http://www.btplc.com/Responsiblebusiness/Supportingourcommunities/Learningandskills/Freeresource
s/AllTalk/default.aspx?s_cid=con_FURL_alltalk
48
See Peter Trudgill: Standard English – what it isn’t, at
http://www.phon.ucl.ac.uk/home/dick/standard.htm
49
Macaulay 1975
non-standard dialect in children who are essentially bi-dialectal. The challenge is to
make sure that children feel as proud of their local variety as they are of Standard
English, and the best contribution that schools can make is to treat them as different
but equal. In the case of grammatical analysis, this means objective comparison:
comparing them as equals, without implying that in some sense the local non-standard
is a poor copy of Standard English. For example, if the local dialect uses was after we
as well as after he, it is different from Standard English, but none the worse for it –
indeeed, the lack of a contrast in the past tense brings the verb to be in line with all the
other verbs, so the local dialect is more regular than Standard English.
The idea that grammar teaching might raise the status of non-standard dialects
deserves a little more discussion. One argument that helped to remove grammar from
the curriculum was that teaching Standard English was inherently prescriptive,
prescribing the forms of Standard English and proscribing non-standard forms. In
prescriptive grammar, we were and those books are correct, and we was and them
books are simply wrong. If this kind of grammar teaching makes pupils feel bad, not
only as speakers but as people, then perhaps all grammar teaching should cease50. The
approach that I am suggesting reverses this argument: grammatical analysis can be
purely descriptive (free of value judgements), so it can be applied as easily to non-
standard dialects as to Standard English and thereby raise the status of the former.
‘Detail’ is my name for the fine linguistic details that children need to know
when speaking and listening. These details include the features that distinguish the
local non-standard from Standard English, but go well beyond them to include any
patterns that are found in more formal or specialised discourse. Many of these are
simply carried over from formal writing, but some are not: more formal ways of
greeting people (e.g. good morning) and addressing them (e.g. sir), ways of
structuring discourse (e.g. by the way) and expressing degrees of certainty (e.g. didn’t
she?), and so on. Grammatical analysis can help children to broaden their range of
constructions in speaking in just the same way as in writing: careful study of recorded
speech not only reveals the new patterns in that sample, but also helps children to
notice new patterns in all their listening. Just as in writing, they have a myriad fine
details to learn before they count as mature competent speakers who can function
comfortably in a wide range of social settings.
However, it is important to recognise the lack of relevant research evidence
here as to how, or even whether, schools can help. It is easy to muster theoretical
arguments for or against the use of grammatical analysis as a tool for expanding
children’s repertoire of linguistic patterns for use in speaking. In its favour, it offers a
plausible way to help children to help themselves by learning from experience, which
is at least more promising than trying to teach all the details directly at school (given
the constraints not only time but also on our collective knowledge of what needs to be
taught). On the negative side, however, we simply don’t know whether it works –
whether a grammatical analysis of a recording, done in class, produces skills that
transfer to ordinary speech. The best I can say is that the idea is plausible, but needs
research.

5.4 Foreign languages

Until the 1960s, the dominant method for teaching foreign languages combined
explicit grammatical rules with translation exercises, so it is now called the ‘grammar-
translation’ method. This obviously presupposed that pupils already knew a
substantial amount of grammar (both concepts and terminology), building on the
50
Trudgill 1975
foundations laid in the English classroom. The method had serious weaknesses, not
least that it treated living languages in the same way as the dead languages, Latin and
Greek, for which it was originally devised, so ordinary conversational skills were low
on the agenda. The problems multiplied when grammar disappeared from English
teaching, so grammar (and translation) fell out of favour, to be replaced (eventually)
by ‘communicative’ methods based on the rather odd idea that learning a foreign
language in a classroom should follow the same pattern as learning ones first
language at home. The argument (encouraged by the fashionable linguistic theory that
language is innate) ran that since small children learned their first language without
the help of grammar, the same would be true for older children learning foreign
languages at school51.
This claim has been explored intensively by researchers, and refuted52. The
research shows that grammatical rules should be taught explicitly, using what is called
‘form-focussed instruction’, rather than left implicit in the hope that learners will
figure them out for themselves. Just as in first-language English, explicit grammatical
analysis encourages both noticing and understanding:
Why is metalinguistic activity [including grammatical analysis] on the part of
learners apparently so valuable? One reason can be found in [the] claim that
while awareness at the level of noticing is necessary for learning, awareness at
the level of understanding will foster deeper and more rapid learning.53
This does not, of course, mean that grammar is all we need, and it is certainly not a
reason to turn the clock back to the 1950s. We also know that learners need high-
quality input and high-quality interactive practice in order to turn this explicit
knowledge into the implicit knowledge that counts as skill in using a foreign
language. But it does mean that grammar can and should play a much larger part in
foreign-language teaching than it does currently.

5.5 Thinking
One of the traditional arguments for teaching grammar was that it was good mind-
training for the elite. A degree in the classics, with the grammars of Greek and Latin
as a major component, used to be considered an excellent training for the higher ranks
of the civil service. This argument is still just as valid as it used to be, except that we
can now generalise it beyond the academic elite. Some experience of grammatical
analysis is probably good for any mind, so long as it is pitched at the right level.
Unfortunately there is very little research evidence to support these claims, but
equally there is no evidence against them.
Grammatical analysis is very similar to mathematics in terms of the mental
demands that it makes. In both cases the learner has to learn a tightly-interconnected
set of concepts such as: noun, subject, object, verb, modifier, adjective; these concepts
all help to define one another through statements such as “A verb’s subject is a noun”,
“A noun may function either as a verb’s subject or as its object” and “A noun may be
modified by an adjective”. In both cases the key concepts are relations between
entities rather than simple entities (e.g. ‘squared’ rather than ‘5’ in mathematics, and
‘subject’ rather than ‘noun’ in grammar). And in both cases it is essential to be able to
apply the analytical system to concrete examples, preferably examples from real life;
so just as the mathematician ‘thinks mathematically’ about some scenario, the
grammarian ‘thinks grammatically’ about a sentence or some other kind of text.
51
Krashen 1982
52
Norris and Ortega 2000, Spada and Tomita 2010
53
Ellis 2008:452.
 Moreover, both mathematics and grammar may be quite abstract, so
mathematicians consider the properties of the (non-existent) square root of minus 1,
while grammarians consider those of the missing subject in a sentence like Come
here! As in mathematics, a grammatical conclusion may lie at the end of a long chain
of arguments and assumptions, so grammatical reasoning can be very challenging;
and any step in the argument may be challenged on either theoretical or factual
grounds. The glory of both subjects is that it is possible – in fact, very easy – to be
wrong, and to be shown to be wrong.
Grammar, then, shares with mathematics the fact that it is a complex, abstract
system that can apply to concrete pieces of real or imaginary experience, and both the
system itself, and its application to a concrete experience, are subject to rational
debate. The most obvious difference between the two subjects is that mathematics is
taught at school, but grammar isn’t. Both subjects can be difficult to grasp and to
teach, so both need teaching methods tailored very carefully to the needs and abilities
of the learners. But nobody would argue that mathematics might be dropped from the
curriculum because of these difficulties or because some learners may not
immediately see their relevance. If we can find ways to teach mathematics to all, why
not grammar?
Indeed, grammar arguably has an even better claim than mathematics on time
in the school timetable. After all, grammar is about the basic organisation of language,
and language is our main tool for thinking and learning. Grammar isn’t an abstract
Platonic system like mathematics that exists ‘out there’; it’s part of our minds, the
product of thousands of hours spent, during childhood, listening to people round us
talking. English grammar is different, in fundamental ways, from French grammar
and Chinese grammar. It forces us to make distinctions, hundreds of times a day, that
French doesn’t force at all – e.g. the difference between it rained and it was raining –
and vice versa – e.g. the difference between tu and vous, or between voisin and
voisine (meaning ‘neighbour’, male or female). Some things are much easier to
express in one language than in another, such as different degrees and kinds of
uncertainty; and the associations embedded in two different languages may be quite
different (e.g. cycling is associated by the English verb to ride with horse-riding, but
by the German verb fahren with driving a car).
This is not to say that our language locks us into an intellectual prison from
which we can’t escape54; on the contrary, language is only one source of influence on
our minds. Another source is, as we all hope, education. Take the distinction which
English forces on us between humans and everything else by making us choose
between the interrogative pronouns who and what. If you ask me: What broke the
window?, I could reply The wind, A falling slate or A bird, but not John – or at least,
not without some comment. In contrast, if you ask: Who broke the window?, I could
blame John but not the wind. This distinction lumps living creatures such as birds
together with inanimate things such as the wind and a piece of stone, in contrast with
humans. But we all know – thanks to education – that this classification obscures a lot
of similarities between humans and other living creatures, and a lot of differences
between either and non-living things like stones and the wind. A more ‘scientific’
classification would cut the cake very differently, so we have two very different
54
This idea of language as an intellectual prison is the strong version of the Sapir-Whorf Hypothesis,
named after the two American linguists who most famously expressed it. Most linguists and
psychologists reject the strong version (Pinker 1994), though it survives in some versions of Post-
modernism. However, a weak version of the Hypothesis, in which language is a major influence on
thinking, though not the only one and not necessarily on all areas of thinking, is widely accepted
Levinson 2003.
‘ontologies’ (ways of classifying things) to choose between, one basically driven by
language and the other by education and science. However worrying this may seem, it
doesn’t seem to be a problem, as every adult lives with both ontologies, and applies
them on different occasions as needed. But this ‘double-thinking’ is something we
should all be aware of, and where would we learn about it other than in grammatical
analysis of English?
Thinking goes beyond mere classification, and the most important questions
are those concerned with how things (or people) are related to one another. Is this an
example of that? Does this cause that? Is this an alternative to that? Did this happen
after that? The more abstract and complicated the reasoning, the more likely it is to
depend on language; and if we want to share it with someone else, we need a
symbolic medium such as a technical diagramming system, or ordinary language.
Seen from this perspective, it is essential for all of us, however academic or non-
academic we may be, to understand this basic tool, so that we are in command of the
tool, and not the other way round. And in particular, it is the general patterns of
grammar rather than the minutiae of vocabulary that we need to understand because
the general patterns are not only more general, but also harder to be aware of.
Fortunately I can finish this rather speculative section with some concrete
evidence that ‘doing grammar’ uses high-level thinking skills. The evidence comes,
once again, from the Linguistics Olympiad in which children as young as 12 years old
struggle to understand how some unfamiliar language works, with grammar
underlying most of the questions. I argued earlier that grammar requires very similar
mental skills to mathematics, and this claim is confirmed by the fact that most of
those who do best in our competition are also studying mathematics; indeed, some of
the champions in the International Linguistics Olympiad are also champions at the
International Mathematics Olympiad.
The next section will explain what it is in grammatical analysis that gives it this
hard, mathematical challenge. However, I shall also explain why it is also a ‘soft’,
humanistic and personal challenge, so that it combines the best of the two very
different worlds of mathematics and literature.

6 What?
What, then, is the grammar teaching that I have in mind? Much as I personally loved
the grammar teaching that was dumped from the curriculum in the 1960s55, the new
grammar teaching is very different. Based on new pedagogy and on new research, it is
interesting, challenging, relevant, highly creative and fun. No doubt the word
hyperbole comes to mind, but I shall now try to justify these ambitious claims.
The following explanation distinguishes two main areas of grammar:
morphology and syntax. Morphology deals with the internal structure of words to the
extent that it reflects the relations among words in the vocabulary; for example, it
recognises the role of affixes such as ing or un which relate word pairs such as run:
running or wire:wiring and tidy:untidy or do:undo. Morphology is really a branch of a
much larger analysis of the entire vocabulary, called lexical analysis, which also
includes patterns such as big:small:size where there is no similarity at all in the word-
forms. In contrast, syntax deals with the relations between words in a sentence,
whether or not these relations are marked by morphology. For example, in the
sentence Cows moo, morphology shows how cows is related to cow and syntax shows

55
See Keith 1990 for a readable and richly illustrated discussion of the dying days of grammar
teaching.
how it is related to moo; and lexical analysis shows how cow/cows is related to cattle,
its ‘hypernym’.

6.1 Lexical relations and morphology

Most graduates know a lot of words, certainly numbering in the tens of thousands and
(depending on how and what you count) maybe in the hundreds of thousands. This is
probably the result of education56, and one of the main aims of education is precisely
to develop a large vocabulary of well-understood and well-integrated words. Each
subject area has its own technical vocabulary which the teacher passes on as a by-
product of the main teaching, but English and foreign languages are special. They do
have some technical vocabulary – not least, the metalanguage of grammar – but their
responsibility includes general-purpose words: ‘academic’ vocabulary in English, and
commonplace vocabulary in the foreign language.
A person’s vocabulary is not just an unstructured heap, but a highly structured
network in which each word is held in place by its links to other words57. The word
hero, for example, is part of heroic and heroine (though this link is clearer in spelling
than in pronunciation), more obviously part of heroes and antihero, and its meaning
relates it to a complex of words such as deed, great, noble, man and admire. In spite
of the homophony with heroine, hero is not directly related to heroin; but it does share
its first syllable with hear.
These links aren’t just a figment of the analyst’s imagination: psychological
experiments show that they really exist in our minds. The main evidence comes from
research on ‘priming’, the effect that an earlier word has when we try to find a later
word in our memory. For example, if you hear the word doctor and then nurse, you
need less time to find nurse (so as to be sure that it’s a word you know already, for
example) than it would if you had just heard the word lorry. The times are measured
in milliseconds, well below the threshold of what we are aware of, but the findings are
extremely robust58. And of course this method allows us to tease apart the different
kinds of link: morphological (hero – heroes or hero – heroic), phonological (hero –
hear), semantic (hero – admire).
Lexical links are the stuff of poetry. Phonological patterns set up expectations
of rhyme and metre, which may be satisfied or disappointed; and at the other end of
language, semantic patterns underlie metaphor, metonymy and association. In
between, we expect phonological similarities to be reflected in morphology, but again
this expectation may be deliberately frustrated. Perhaps the main point of poetry is to
create expectations, on the basis of normal language, and then to deliberately frustrate
them. Linguistics can help to understand and explore the expecations and frustrations,
and then to hand over to the reader or critic to puzzle out the poet’s intentions or
effects.
For example, a limerick has a very strong pattern of strong and weak beats
which are easily mapped onto the more complex pattern of ordinary English metre, as
in this anonymous example59:

The limerick packs laughs anatomical

In space that is quite economical.
But the good ones I've seen
56
See http://en.wikipedia.org/wiki/Vocabulary#Native-language_vocabulary for a brief review.
57
Aitchison 1994
58
See any introductory textbook on psychology, such as Reisberg 2007
59
From en.wikipedia.org/wiki/Limerick_(poetry)
 So seldom are clean
And the clean ones so seldom are comical.
The standard metre for a limerick seems to be a matter of dispute, although they are
very easy to recognise and the metre is easy to describe informally. A limerick
assigns each line a light upbeat followed by either three or two units which we may as
well call feet, with each foot consisting of a strong syllable possibly followed by some
weak ones; so the metre builds on the alternation of strong and weak syllables. Here is
the same example with a metrical analysis shown by the underlined strong syllables
and a vertical bar at the start of each foot:

The | limerick packs | laughs ana|tomical

In | space that is | quite eco|nomical.
But the | good ones I've | seen
So | seldom are | clean
And the | clean ones so | seldom are | comical.

What can a lexical analysis say about this poem? The obvious things relate to
pronunciation. Most of the foot boundaries lie between words, but those at the end of
the first two lines cut straight through the words anatomical and economical; and
similarly, although these two words rhyme with each other, as well as with comical in
the last line, this similarity does not indicate a similar morphological analysis, in
which anatomical can be dissected into ana plus tomical; on the contrary, its
morphology consists of anatomic plus al (or, possibly, anatom plus ical). Both the
metre and the rhyme invite the reader to relate tomical, nomical and comical – but
there is no underlying relation.
Then there is meaning. The combination laughs anatomical makes no direct
sense either syntactically (compared with anatomical laughs) or semantically – how
can a laugh be anatomical? However it makes complete sense in this context by
hinting at the familiar phrase belly laughs, and it is a fair guess that a psychological
experiment would show that laughs anatomical does indeed prime belly. The fun
comes from a kind of competition in which the author challenges us to decode his
message in spite of the barriers he erects, and we succeed.
The semantic links to laughter continue with good ones. As in normal English,
ones refers back to a previous noun, in this case limerick; and as in normal English,
good could be called a chameleon word because its meaning varies wildly with
context. What makes a knife ‘good’ is very different from what makes an exam result
or a person ‘good’, so we have to understand good in this case as funny. This process
of interpretation follows the normal rules of English, but it allows the author to avoid
using the word comical or funny in line three. The monosyllable good fits the metre
better than either of these polysyllables, and reserves comical for the punchline at the
end. The semantic link between good and comical is important because this pair are
contrasted with the word clean, which is repeated directly.
The example hardly deserves such detailed analysis, especially when
compared with the great masterpieces of poetry, but it has the merit of brevity and
accessibility. The main point that I want to make clear is that lexical analysis has a
great deal to offer in the close reading of texts. You might object that this discussion
has taken us well beyond the usual confines of grammar, but this is not so. Even if
you define grammar so that it consists of nothing but morphology and syntax, it has to
make contact with the other parts of language, including pronunciation and meaning,
so when considering either of these areas, and especially when considering them
together, we are in effect doing grammar.
I’m not sure whether limericks count as literature, but it doesn’t really matter
because lexical analysis can be applied to any kinds of texts, whether literary or non-
literary. Close reading of any professionally written text is a good way to learn the
writer’s trade; and of course the same is even more true in a foreign language. Close
lexical analysis can even be applied to spoken conversation. But textual analysis of
texts is not the only way to use lexical analysis in an English or foreign-language
classroom. After all, lexical relations between words exist whether or not the words
are used in some text, so it makes good pedagogical sense to consider them as a
‘system’ – a network of interconnected elements. Indeed, I shall argue in section 7 in
favour of considering elements as part of the system before looking for them in texts.
What, then, is this system of vocabulary? In other words, how do we organise
words in our minds? One thing we can be sure of is that it is very, very large, with
hundreds of thousands, possibly millions, of elements and interconnections. In case
this strikes you as a wild exaggeration, let me take you through some plausible
figures. Suppose that you know 50,000 words60, that each of these words, on average,
has two meanings61, that each word is, on average, spelt with 4 letters62 and
pronounced with six phonemes63. This means that each word is linked to twelve
(2+4+6 = 12) separate elements, which immediately gives us 600,000 links. Each link
takes us to another element – a meaning, a phoneme or whatever – so we need to add
these in, but obviously the same small inventory of phonemes and letters is shared by
all words, so the main addition are the units of meaning – say, 100,000 of them,
though synonymy will reduce the figure somewhat. That takes the total to 700,000,
without saying anything at all about syntactic links to other words and word patterns
that combine with each word – e.g. the link between hero and romantic, or the one
between belly and laugh which I mentioned above. I have no figures, but I should be
very surprised if they did not take us well beyond a million. And even then I haven’t
mentioned all the contextual associations and style restrictions and so on and on – the
stuff of much English teaching.
This simple calculation has enormous implications for education, because it is
education that turns a five-year old’s 5,000 words into the educated adult’s 50,000,
instead of the much smaller number that we might expect from growth at the rate of
1,000 per year gradually reducing as the child ran out of words to learn. How does
education achieve it? Once again, one thing is very clear: there simply isn’t enough
time in the school day to teach all that vocabulary, item by item. This is why
‘noticing’ is so important, because it is only by noticing new words and new details of
existing words encountered throughout the day that the child can develop so far and
so fast.
Another certainty is that our vocabulary is a network, not a list. This certainty
comes from the psychological experiments mentioned earlier in which an earlier word
primes a word that follows it in the experiment if the two are linked in any of a
number of ways – by meaning, by pronunciation, and so on. These links must be links
in a network which allow a word to be ‘next to’ all sorts of other words or concepts at
the same time. This is a radically different kind of organisation from what we find in
our dominant model for vocabulary structure, the typical dictionary. A dictionary is a

60
Aitchison 1994:7
61
Levickij and others 1999
62
http://hearle.nahoo.net/Academic/Maths/Sentence.html
63
Cutler and others 2004
list, organised alphabetically, and not a network. A much better model for thinking
about ‘the mental lexicon’, as our vocabulary is often called, is the internet, where any
page may cross-link to any other page. In short, our vocabulary is much more like the
network in Figure 1 than the dictionary.

Figure 1: Vocabulary as network or list?

Any network structure is too complicated to describe in ordinary prose so
subjects that deal in networks tend to develop a standard notation. Chemistry has
molecular models, geography has maps, music has the five-line staff. Linguistics has
notations too – plenty of them – and you may find some of them useful. For
vocabulary structures we show words and their meanings on different ‘levels’, with
the meanings above the words, and we put morphemes on another level below the
words, with phonemes and written letters on an even lower level. The idea of levels is
a metaphor, based on the idea that more abstract things are ‘higher’ than more
concrete things. This allows us to distinguish different kinds of elements clearly while
also allowing us to relate them with lines. We also have a notation for distinguishing
different levels when naming their elements: words are in italics, morphemes are
between curly brackets, phonemes are in the slants that you may have seen in phonics
material, written characters are in diamond brackets and meanings are between
inverted commas. When we combine these two notations we get a diagram like Figure
2 (in which I have used the UK government’s notation for phonemes64 rather than the
linguisticians’ International Phonetic Alphabet).

64
http://tinyurl.com/nc2012english
 level

‘horse’ ‘croaky’ meaning

steed horse hoarse word

{steed} {horse} {hoarse} morpheme

/steed/ /haws/ phoneme

<steed> <horse> <hoarse> letter

Figure 2: A fragment of a vocabulary network

What the network idea means for education is that learning takes place, not by
adding one item after another to an open-ended list, but rather by enriching the
existing network. In fact, arguably it is impossible to learn anything which has no
connections at all to existing knowledge, because connections are what we create
when we understand something and we can’t learn something that we don’t
understand at all. Learning and understanding clearly go hand in hand, so the more
links tether a new concept to the existing network, the more robust that new concept
will be. This is as true of a new word as it is for a new meaning, and it is as true of
words in foreign languages as it is in English. Learners need to actively search for
links to existing vocabulary, and grammatical analysis can help by guiding this search.
Which brings us to morphology. This is the study of how the form of a word –
its pronunciation and/or its spelling – may relate it to other words as a signal of
semantic (or other) relations. For example, the morphological structure of untidy
relates it, in one direction, to unnecessary and all the other words containing negative
un, and also, in the other direction, to the adjective tidy (and ultimately to the verb
tidy and to tidiness, tidier and tidiest). The word-parts are called morphemes (which
gives me an opportunity to demonstrate the benefits of morphology in tethering a new
word to the existing vocabulary: think phon-eme, graph-eme; morph-ology, Morph65,
meta-morph-osis).
One of the benefits of morphology is to allow new words to be created – an
example of the creativity that I alluded to earlier. How many words can you make by
combining the morphemes un, re, pre, ness, er and ed with tidy? We’ve already had
untidy, but how about retidy and pretidy? In these cases we’re dealing with the verb to
tidy, but it still consists of the same morpheme tidy. Then we have tidiness and
untidiness, tidier and untidier, and tidied and untidied. (See if you can work out which
of these are based on the adjective tidy and which on the verb.) But tidied can just
65
http://en.wikipedia.org/wiki/Morph_%28character%29. Morph was made of plasticine so he could
easily change shape.
about be an adjective (as in his freshly tidied bedroom), so we can also have untidied
(with two completely different meanings, depending on whether it’s based on the verb
to untidy or the adjective tidied), and maybe even untidiedness. Then there’s the other
meaning of tidier as a noun based on the verb to tidy, and likewise for untidier; not to
mention retidier and pretidier. And how about repretidy and preretidy, or even
repretidiedness? Your vocabulary network may be much less fixed than you think, and
much more open to creative innovation. Word-creation is an important skill for
novices whose vocabularies are still growing fast, but it is also fun.
Word analysis is part of grammatical analysis, and can be fun too. What is the
longest real English word? If word-creation is as creative as I claim, there can be no
such thing, but an urban myth says it’s antidisestablishmentarianism. Well, how’s this
made up? Let’s deconstruct it by gradually stripping away the morphemes, giving:
antidisestablishmentarian
antidisestablishmentary (note the y to i rule in reverse)
antidisestablishment
disestablishment
disestablish
establish
This is the only possible order; for instance, we couldn’t recognise establishment
without dis, because dis only combines with verbs (such as establish), and not with
nouns. This useful exercise is an excellent way to make children aware of lexical
relations and morphology. Once again, a notation is useful in order to save repeated
rewritings. Linguists tend to use brackets, like this:
[[[anti[[[dis[establish]]ment]]ari]an]ism
But when brackets pile up they can be unhelpful, so you may prefer to use
underlining, as in Figure 3.

antidisestablishmentarianism

Figure 3: Analysis of ‘the longest word in English’

Perhaps the most important thing about lexical relations and morphology is the
strong focus on relations – relations between the parts of a word, but even more
importantly, relations between a word and its ‘neighbours’ in the network of
vocabulary. These relations are multiple and varied, so it is vital to keep the different
kinds of relation distinct; for example, the word horse is related in quite different
ways to the word steed, the idea ‘horse’ and the morpheme {horse}. We even have
readily available names for these relations: in relation to horse, steed is its synonym,
‘horse’ is its meaning and {horse} is its ‘realization’ (a technical term, but easy to
understand if you think of the relatively concrete morpheme making the relatively
abstract word more ‘real’). Sorting out this conceptual tangle, with the emphasis on
relations, is a really challenging and productive mental exercise, quite apart from any
benefits it may have for developing language skills. We’re a long way here from the
so-called ‘grammar grind’66 of pointless and mindless exercises on minutiae.

66
The phrase grammar grind is a misquotation based on Robert Browning’s poem about a specialist in
Greek grammar, “The grammarian’s funeral”: ‘So, with the throttling hands of death at strife, / Ground
he at grammar; ...’. (Crystal and Crystal 2000:117)
6.2 Syntax
The same focus on relations is repeated in syntax, but even more so because syntax is
about how words are related to each other when combined into sentences. This is
where we talk about subjects and objects, about modifiers and heads – all the names
of relations. But whereas lexical relations are in the system – the lexical network –
those of syntax are in the text. In some ways this makes syntax easier; after all, if I ask
you to analyse the syntax of the sentence The cow jumped over the moon, you know
exactly which words need to be related: those in the sentence. This is easier than
thinking about a single isolated word in order to work out which other words it might
be related to (e.g. as a synonym). On the other hand, it also makes syntax harder
because, ultimately, every word needs a place in your analysis, however tricky that
place may be. The good news, however, is that, if they want, novices can do the easy
words now and leave the harder ones till later, as I shall explain below.
As with lexical relations and morphology, syntax has two faces. One face
looks at the text, and the other at the system. Take The cow jumped over the moon, for
example. When facing the text, syntax allows us to analyse its structure; for instance,
it lets us show both that the and cow belong together, and also how they are related to
each other. But when facing the system, syntax tells us that words like the can
combine with words like cow, and how they combine (e.g. in what order and with
what effect on meaning). Like the two faces of a coin, neither face can exist without
the other; so we only know that the and cow belong together in this text because we
know that the system allows them to combine in this way. As native speakers of
English we tend to take the system for granted, though it is always there in the
background and always deserves some exploration; for us, the main challenge is
working out how the sentence ‘works’. But when studying a foreign language it is the
other way round: we’re still learning the system, so any bit of text is an opportunity to
develop our knowledge of how the system, rather than the sentence, ‘works’.
Syntactic relations are controversial because there are two different traditions,
separated geographically and historically, and each offering a different approach. As
so often happens, it may be that both are right to some extent, but my sympathies will
become very clear below. I’ll illustrate the different approaches with a very simple
sentence, Small babies often cry.
First, in terms of history, comes the ‘dependency’ approach, which is based on
direct relations between individual words. For this approach, there is a relation
between small and babies, and another between babies and cry; but (crucially) small
babies does not form a unit. These relations are called ‘dependencies’ because they
subordinate one word to another, so that the former depends on the latter; so small
depends on babies, and babies on cry. In almost every case, the dependent modifies
the meaning of the other word by making it more precise, so small babies is more
precise than babies would be on its own. Similarly, often also depends on cry. This is
a grammatical tradition that goes back at least to the Arabic grammarians of the 8th
century (who developed sophisticated ways of analysing Arabic grammar in what is
now Iraq)67, and that has provided the basis for the teaching of grammar in some
countries to this day.
One of the main contributions of this tradition is a notation for showing the
syntactic structure of a sentence which has come to be called ‘sentence diagramming’,
as in Figure 4.

67
Owens 1988
Figure 4: A Reed-Kellogg diagram
Invented in the middle of the 19th century by two Americans, Alonzo Reed and
Brainerd Kellogg, this system is still taught in some schools in the USA68. Indeed, it is
still so popular there that someone has created a website where Reed-Kellogg
diagrams are computer-created to order69. (This is how Figure 4 came into being.)
Various people have told me that the same system is also used in some countries in
Europe. The diagram uses the vertical dimension to show subordination, so small and
often are each dependent on the word above them. Because the direction of
subordination between babies and cry is less obvious, the diagram puts them on the
same level. One major drawback of these diagrams is that they don’t reflect the actual
order of the words, so there is nothing to show whether often stands before cry or after
it.
A number of other notations have been developed for dependency analysis, but
I have to declare an interest here, because I have my own70, which (hardly
surprisingly) I think is the best available. It has a number of advantages over its rivals,
not least being the fact that you don’t need to rewrite the examples because you’re just
adding arrows between words. This makes it ideal for use in the classroom. Figure 5
shows how it handles our example. All you need to know is that arrows point at
dependents, and dependencies can be classified by labels such as ‘subj’ (for ‘subject’).
The arrow pointing straight down at cry shows that this word does not depende on any
other, so it is the ‘root’ or anchor on which all the other words depend.

sub
j

Small babies often cry.

Figure 5: A Word-Grammar analysis

The rival to dependency analysis is called ‘constituency’ analysis. This was
invented in the early twentieth century by an American linguist, Leonard Bloomfield,
but it was then taken up by the world’s most famous linguist, Noam Chomsky, who
called it ‘phrase-structure’ analysis71. The idea behind this analysis is that the
fundamental relation is that between a whole and its parts, rather than a direct relation
between those parts; so the analysis shows how a sentence can be divided into its
major parts (called phrases), each of which may be further divided into smaller
phrases, and so on until no more parts can be identified. The standard notation for this

68
http://www.polysyllabic.com/olddiagrams
69
http://1aiway.com/nlp4net/services/enparser/
70
http://www.phon.ucl.ac.uk/home/dick/wg.htm
71
http://en.wikipedia.org/wiki/Phrase_structure_grammar
kind of analysis is the ‘tree’, illustrated in Figure 6, in which lines show how small
babies forms one unit and often cry another.

Small babies often cry.

Figure 6: A phrase-structure analysis

When applied to easy examples there is little to choose between these
approaches, but most examples that merit analysis are interesting precisely because
they are complicated, and at that point the advantages of the dependency approach
become clear. For example, in Babies keep crying, there are good reasons72 for
thinking that babies is the subject not only of keep but also of crying. (Grammarians
generally call this pattern ‘raising’.) A phrase-structure analysis doesn’t allow
branches to tangle, so babies can’t combine directly with crying and some kind of
special device has to be found for howing this link, such as an extra ‘empty’ branch in
the tree linked to babies. In contrast, a dependency analysis does allow it, provided
we distinguish dependencies that do affect word order from those that don’t (e.g. by
drawing the latter below the words). The two kinds of analysis are contrasted in
Figure 7.

sub
j
Babies keep crying. Babies keep crying.
sub
j
Figure 7: Dependency and phrase structure analyses of raising.
At a research level, it may turn out that we actually, i.e. mentally, recognise
both kinds of structure73. This wouldn’t be surprising if you think of social structures,
where we recognise not only direct ‘dependencies’ between individual people, but
also some larger ‘phrases’ such as families and institutions, of which individuals are
just parts. However, for most practical purposes such as teaching, dependencies are all
you need.
One of the attractions of dependency analysis is the possibility of partial
analyses, which allow you to focus on just one relation without working through the
entire sentence’s structure. This is helpful when dealing with syntactic ambiguities –
sentences that allow two or more analyses, each with a different meaning. Clearly
these are best avoided in writing, so it is important for novices to be aware of them. A

72
See, for example, the artice on ‘raising’ in
http://www.phon.ucl.ac.uk/home/dick/enc2010/frames/frameset.htm.
73
Tallerman 2009
good way into teaching about syntactic ambiguities is through jokes and other
deliberately constructed examples. Take, for example, the following:
Policeman to little boy: “We are looking for a thief with a bicycle.” Little boy:
“Wouldn’t you be better using your eyes?”
The humour lies in the different structures that the policeman and the boy assign to
the policeman’s utterance, where the key difference lies in the word with: does it
depend on thief (giving the meaning, ‘a thief with a bicycle’), or on looking (giving
‘looking with a bicycle’)? Figure 8 gives the two partial analyses needed:

We are looking for a thief with a bicycle.

We are looking for a thief with a bicycle.

Figure 8: A syntactically ambiguous sentence

To see the reality of syntax in your mind, consider a special case of ambiguity
which is called the ‘garden-path sentence’ because it leads the reader up the garden
path. The first interpretation remains the obvious one right up to the last word, which
forces a completely different one. Here’s a simple example74, starting with the words:
Fat people eat .... What do you expect now? Maybe too much? That’s obviously right
– until you see the actual last word of the garden-path sentence: accumulates. For
most readers, their expectations are so strong that the experience of reading this word
could almost be described as painful. If you want to understand why rethinking the
structure was so painful, it may be helpful to draw a picture of what you were
expecting, followed by the structure you had to put in its place – a complete
reorganisation.

Fat people eat ........ Fat people eat accumulates.

What about classifying words? You may have noticed that I haven’t even
mentioned the so-called parts of speech (which I prefer to call simply word classes,
because that’s what they are): noun, verb, and so on. This is deliberate, because I
don’t want to encourage the idea that classifying words is not only the beginning of
syntax, but also the end. Word classes are important because they allow important
generalisations in the system, but when applied to particular words in particular
sentences, they don’t really help much, and especially not if the aim is to understand
how words convey meaning. For instance, my garden-path sentence switches the
word class of fat from adjective (as in fat people) to noun (as in fat accumulates), but
the crucial change is in the relations between the words: fat depends on accumulates
instead of on people. An analysis that simply changed fat from adjective to noun
would miss the point of the example almost entirely.

74
From http://www.fun-with-words.com/ambiguities.html, a wonderful treasurehouse of examples.
 Nor have I suggested recognising ‘clauses’, in spite of their major role in
traditional grammar. A clause is simply a verb plus all the words that depend on it
(directly or indirectly), and it is imporant simply because it’s a potential sentence; but
in dependency analysis all we need to recognise is the verb and its dependencies.
Saying ‘This is a clause’ actually adds nothing to what you have already said if you
say ‘This is a verb, and all these words depend on it.’ And if no clauses, then also no
subordinate clauses or coordinate clauses, and no so-called simple, complex and
compound sentences (distinguished by the respective presence of no clause, a
subordinate clause and a coordinate clause). If you recognise those terms, you will
probably know that they are very hard to apply and not particularly illuminating, so I
hope you will welcome the invitation to abandon them in your teaching. If you don’t
know what I’m talking about, so much the better.
What, then, do you need to know in order to teach syntax? More precisely,
what do you need to know in order to feel confident in teaching syntax? That means
feeling that you can say something sensible about any random example that comes up
in class, as you probably could if the question was literary. When I taught English
grammar to first-year undergraduates, I found enormous variation in how easy they
found it. For some, it was easy and largely a matter of common sense; and by the end
of the ten-hour course they could analyse almost every word in a 100-word text. But
for others, it was a struggle and clearly not something that came at all naturally. Given
what I said in 5.1 about ‘noticing’, I think this is exactly what we might expect. Those
who can notice words, as words, find it easy; but some people simply can’t see the
words for the meanings. For instance, if they look at accident, all they see is an event,
and it is only with great difficulty that their minds can separate the event from the
word that expresses it so as to classify this word as a noun (in spite of its abstract
meaning). Maybe if those undergraduates had done serious grammar in primary
school they would have learned to notice words; but they hadn’t. We are where we
are, so some people have to work much harder than others at building confidence in
grammar.
For those who find it easy, the dependency structure of a sentence is mostly
obvious because it follows the meaning. There are a few tricky points where it’s
obvious that there is a dependency, but it’s less obvious which word depends on
which. For example, which depends on which in the pair is sleeping? There are very
good reasons for saying that sleeping depends on is75, but many people find this quite
counterintuitive – and even more so with a popular analysis of the cow in which cow
depends on the76. As long as you’re consistent it doesn’t much matter which analysis
you choose – and you’ll probably find some grammarian who supports your analysis.
But even those who struggle will certainly become more confident with practice. I
return to this issue in 8.
It should be clear that grammatical analysis (whether lexical or syntactic) is
just as relevant to non-standard dialects as it is to Standard English. And similarly, it
is just as relevant to literary language as to non-literary. To combine these points, I’ll
suggest what grammatical analysis might contribute to a reading of a classroom
classic, John Steinbeck’s Of Mice and Men77. Here is a more or less randomly chosen
passage (p. 32):

75
See ‘Auxiliary verb’ in http://www.phon.ucl.ac.uk/home/dick/enc2010/frames/frameset.htm
76
See ‘Determiner’ in http://www.phon.ucl.ac.uk/home/dick/enc2010/frames/frameset.htm
77
Penguin Books, 2006
 George stared at his solitaire lay, and then he flounced the cards
together and turned around to Lennie. Lennie was lying down on the bunk
watching him.
“Look, Lennie! This here ain’t no set up. I’m scared. You gonna have
trouble with that Curley guy. I seen that kind before. He was kinda feelin’ you
out. He figures he’s got you scared and he’s gonna take a sock at you the first
chance he gets.”
Lennie’s eyes were frightened. “I don’t want no trouble,” he said
plaintively. “Don’t let him sock me, George.”

My first comment states the obvious: George and Lennie don’t speak Standard
English. And yet, given the examples in this passage, any British class could easily
create more dialogue in George’s dialect because the non-standard features are
familiar from British non-standard dialects, or even from casual standard speech.
Many British children (possibly most of them) would use ain’t for isn’t, no for a, seen
for saw78, and gonna for going to, and would use like where he uses kinda.
One particularly interesting bit of non-standard feature is the lack of are in
you gonna, especially when compared with he’s gonna. Why does the verb be come
and go like this? One easy explanation is that it’s simply a matter of pronunciation:
are would in any case be reduced to ’re, which is so weakly pronounced that in some
cases it simply disappears. Like most other Brits, I do the same when I ask the
question You gonna be long?, but not in He gonna be long?, where I would always
pronounce the is to give Is he gonna be long? Something similar may be happening in
the sentence I seen that kind before, with have reduced to ’ve and then to nothing at
all It is important that children understand how non-standard dialects work, and
especially so when that is what they themselves speak at home. Studying non-
standard dialect in literature may be a good way to neutralize the discussion so that
non-standard speakers don’t feel personally threatened.
Another noteworthy syntactic feature of this passage is the way it
distinguishes George from the simple-minded Lennie. Not only does George do most
of the talking, but he even uses more complicated syntax. Notice in particular his take
a sock at you which Lennie picks up but simplifies to sock me.
Just to repeat my earlier point, the main reason for doing this kind of close
reading of the syntax is to encourage the pupils to notice syntax, on the assumption
that if they notice it, they can learn from it and maybe learn to write as well as
Steinbeck.

7 How?
In 1999, the Qualifications and Curriculum Authority published a collection of papers
called Not whether but how. Teaching grammar in English at key stages 3 and 4, and
in 2000 the Standards and Effectiveness Unit published a 200-page book called
Grammar for writing, aimed at key stage 2 pupils in primary schools. Both books
started from the point in the debate about grammar that I have now reached here: I

78
Note that George’s I seen that kind before may be the non-standard equivalent of I saw that kind
before, rather than of I’ve seen that kind before. The latter is what we would expect in British English
because we use the present perfect (have seen) rather than the simple past (saw) for events that have
‘present relevance’. However, the rules are different in Standard American English, so they may also be
different in George’s dialect. In SAE, present relevance also allows the simple past where there is
present relevance, as in I lost my key. Can you help me look for it? For more on this difference, see
http://esl.about.com/od/toeflieltscambridge/a/dif_ambrit.htm.
hope to have established that grammar should be taught, and what this teaching
should include. But that’s only half of the debate, and the easy half at that. Perhaps the
main problem with grammar teaching before it died was the method of teaching,
which probably turned off teachers even more than pupils.
Modern pedagogy can do better. But this is just an aspiration until we have
quite concrete ideas of how it translates into schemes of work and classroom
activities. And in spite of these two excellent books, both full of good ideas, thirteen
years later nobody could claim that grammar is satisfactorily embedded in our
classrooms, either primary or secondary. Moreover, many of the issues that they
raised are still being debated, so there is still room for ideas. What I hope to offer here
is a reasoned set of principles that integrate the ‘how’ with the ‘why’ and the ‘what’ in
a way that wasn’t possible in either of those books.
The most important pedagogic consequence of my argument so far is probably
that grammar must be taught systematically. This is partly because grammatical
concepts are so tightly related to one another. For instance, to properly explain what
modal verbs are you have to be able to talk about verbs, auxiliary verbs, suffixes,
subjects, complements, and finiteness; and each of these concepts in turn builds on a
range of other concepts. You can, of course, help a lot of pupils simply by giving
examples: “Modal verbs are verbs like must, will, may and should”; but some pupils
may not be able to see what makes these verbs different from others so that your
examples leave them struggling to understand. But another reason why systematic
teaching is so important is that grammar serves many different purposes – not just
writing and reading, but also speaking and listening in English, plus foreign-language
learning and general mental development. This being so, the teaching of grammar has
to follow its own logic rather than that of another area of study. Both of these
arguments for the systematic teaching of grammar can no doubt be matched by
arguments for systematic teaching of basic arithmetic.
This principle is already accepted in Not whether but how, but there is an
unresolved tension between it and another, that grammar should be integrated. Here
are two relevant bullet points from Not whether but how (p.6):
 Explicit grammar teaching should be integrated into the overall English
curriculum.
 Systematic planning should ensure progression and development over time.
I would add that integration is also possible with the overall curriculum for foreign
languages, but the main question is how tightly the teaching of grammar should be
integrated with other areas of study. A popular view is that grammar should only be
taught ‘as needed’79. But as needed by whom, and for what purpose? What if one
student needs to know about modal verbs but the rest of the class doesn’t? What if
modal verbs are needed for learning German, but not for English? What if modal
verbs are important now, but aren’t needed again for another two years? What if a
need arises unexpectedly, and the teacher has had no time to plan the relevant bit of
grammar (or, worse still, to learn it)? The idea of teaching grammar only as needed,
and only when it can be immediately integrated into some teaching activity, sounds
attractive, but actually makes very little sense in practical terms.
This discussion seems to lead to an impasse: on the one hand, grammar has
its own powerful internal logic, so it should be taught according to that logic; but on
the other hand, it should be tightly integrated with other activities (such as learning to
write English or to speak German) which also have their own internal logic. How can

79
Weaver 1996
we reconcile these two competing demands? If we follow grammar’s own logic, we
have a separate grammar lesson every week, and no connections with anything else;
but if we teach it only when needed for some other kind of learning, we lose the
grammatical plot. I believe this dilemma has bedevilled the teaching of grammar for
too long, and its horns need to be grasped.
The solution seems obvious: a compromise based on the idea of a ‘spiral
curriculum’ as defined by Bruner80:
A curriculum as it develops should revisit this basic ideas repeatedly, building
upon them until the student has grasped the full formal apparatus that goes
with them.
Given the complexity and abstractness of grammar, the spiral syllabus is ideally suited
to learning it. For example, according to the 2012 draft National Curriculum for
English in primary schools, the child learns to recognise verbs in Year 1 in connection
with spelling various suffixes and the prefix un, then in Year 2 verbs come back into
focus in relation to tense and the consistent choice of tense in a narrative. Year 3
returns to verbs in a discussion of the choice between simple past and present perfect
(e.g. saw or has seen) and they reappear in Year 4 in connection with differences
between Standard and non-standard verb forms (e.g. I saw and I seen). Year 5 talks
about how to make verbs from nouns or adjectives (e.g. privatise), and about how
modal verbs express degrees of uncertainty; and Year 6 includes work on passive
verbs. This is a perfect spiral curriculum because each year returns to the same basic
concept – ‘verb’ – and enriches it a little more so that it is that much more robust and
ready for the following year to build on. This idea of enriching existing concepts is
exactly right if knowledge really is a network, as I suggested in 6.1; and once one
concept, such as ‘verb’, is well established, it can be used to enrich other concepts
(such as ‘subject’ or ‘modal verb’).
Conversely, of course, a concept or term introduced in one year will wither
and disapear (at least for all practical purposes) if it is not revisited in later years; so
there is no point in ‘doing grammar’ without a master plan for the whole school which
guarantees a revisit.
The spiral curriculum has implications for terminology as well. There is a
powerful urban myth about grammatical terminology being difficult, and perhaps a
distraction from the business of teaching the concepts of grammar; and because of this
myth, there has been a tendency ‘to teach grammatical understanding or concepts
without reference to the relevant terminology’81. The idea of terminology-free
grammar is still popular, not only among teachers but also among language courses
for adults, which are often promoted like this advertisement for ‘the business of
speaking Spanish without grammar labels, textbook exercises or confusing
terminology’82. Some teachers even believe that they help students by making up their
own private terminology. The flaws in this myth are obvious: grammatical
terminology is no different from terminology in any other field in terms of either
intrinsic difficulty or quantity, and nobody suggests that any other field should be
taught without terminology. Imagine teaching geography without talking about
latitude and longitude, or mathematics without square roots and multiplication, or
English literature without metaphor and stanza. Moreover, the debate is especially odd
in connection with grammar, of all subjects, because grammar is inseparable from

80
Bruner 1960:13
81
Myhill 1999:9
82
http://shortcuttospanish.com/blog/welcome-to-my-spanish-shortcuts
vocabulary and everybody in education accepts the importance of vocabulary in
acquiring new concepts.
Far from feeling embarrassed about grammatical terminology, teachers
should use it as much as possible so as to help children to enrich the concepts behind
the terms. This means that different teachers need to use the same terms for concepts.
If the Year 1 teacher talks about ‘doing words’, pupils may not access this concept
when their teacher in Year 2 talks about ‘action words’, and similarly for ‘verb’ in
Year 3. Indeed, a consistent technical terminology is a prerequisite for any spiral
curriculum, because it is only the shared terminology that links one ‘visit’ to the next.
The same need for consistency applies to the teachers of English and of foreign
languages.
But how can teachers ensure that terminology is consistent? This clearly
calls not only for a school-wide policy on grammatical terminology, but also for a
nation-wide one because of the transition from primary to secondary. Rather
obviously, the spiral curriculum breaks down if the secondary school uses different
terminology from the primary school. The good news is that grammar does, in fact,
have a rather rich set of very well established terms, entirely comparable in ancestry
with those of other ancient subjects such as mathematics. If a term such as ‘verb’ has
been around since the Latin grammarians, it seems a great pity not to use it in the 21st
century. Even better news is that the draft National Curriculum for English includes a
glossary of recommended grammatical terms, so there is no longer any reason for
different teachers or schools to use different terminology.
Terminology is essential for making grammatical analysis explicit, but it
isn’t the only tool we have. Another is the use of diagrams using the various notations
that I discussed in section 6 – networks for lexical relations, brackets or underlining
for morphological structure, arrows or trees for syntax. The challenge here is not to
establish uniformity, as with terminology, but to establish any kind of use, because so
far as I know virtually no teacher uses diagrams for grammar. This is a great pity
because diagrams are an essential tool for helping novices to cope with the
abstractness of grammar, and especially of grammatical relations; a visible arrow
linking two words is a great way to show that these words are related. I hope my
arguments will persuade some teachers to try diagrams.
I have argued so far for three basic principles of grammar teaching:
 that it should follow a spiral curriculum in which basic concepts are revisited in
successive years and in different subjects so that they can be gradually enriched,
and made available for defining other concepts.
 that it should use the standard technical terminology of grammar so as to ensure
continuity from teacher to teacher and from school to school.
 that it should use diagrams in order to make grammatical relations concrete.
I have also accepted the principle that it should be integrated with other activities, but
the spiral curriculum means that the activies concerned can be selected according to
the child’s existing knowledge of grammar, which is rather different from the idea of
teaching grammar only ‘when needed’.
Having established these general principles I can now consider how they
can be translated into effective teaching activities. Here all the usual pedagogical
principles apply: that the teaching should be focused on a clear purpose, and that the
best way to achieve that purpose is to engage the pupils in enjoyable activity which
they can relate to their own lives. The units of Grammar for writing already satisfy
these criteria well, so I shan’t try to add to them. However, this collection of units is
aimed exclusively at primary schools, so I shall now suggest some grammar-based
activities for a secondary English classroom.
One is the activity called ‘sentence combining’ which I mentioned in 5.1, in
which the teacher provides two or three short sentences which the class then combines
into a single longer one. With pupils working in groups, the likely outcome is that
different groups will combine the sentences in different ways, so the activity provides
an opportunity for discussing the effects of the various solutions. The activity can be
done without using any grammatical terminology, and even without terminology there
is good research evidence for improvements in the children’s writing skills83.
However, it seems likely that these improvements could be increased by using
terminology to make the learning more explicit. For instance, they could simply note
that it is possible to combine He likes apples and She likes pears either to give He
likes apples and she likes pears or He likes apples whereas she likes pears. But they
learn much more if they distinguish these two combinations as coordinate and
subordinate, especially if they then note that the subordinate (but not the coordinate),
can be moved to the front: Whereas she likes pears, he likes apples. One of the
positive features of sentence combining is the creativity needed in order to find a
plausible way to combine sentences, especially as possible connections becomes less
obvious.
Another activity is the study of jokes, which often depend on grammatical
ambiguities including the special ambiguities found in the ‘garden-path’ sentences I
discussed in 6.2. Puns are jokes which are usually based on homonymy (two words
sharing the same sound or spelling), which are relatively easy to spot given the lexical
network we have in our minds. But the most impressive puns go well beyond any
lexical connections by building on the chance juxtaposition of two words, as in the
old saying: It’s not the cough that carried him off, but the coffin they carried him off
in. The class may have views on how funny this is, but those of us who love it are
impressed by the skill needed to get two words, off in, to rhyme with one, coffin,
especially since these two words are unrelated in the network, and only incidentally
related in syntax. This claim calls for the syntactic diagram in Figure 9 to show that
both off and in depend on carried (giving carried off and carried in (the coffin)):

It’s not the cough

that carried him off,
but the coffin

they carried him off in.

Figure 9: The syntax of a pun

Perhaps the most familiar grammatical pun, at least for a linguistician, is also a
garden-path sentence: Time flies like an arrow; fruit flies like a banana.84 This cries
out for a diagram to explain the gross syntactic difference between the two clauses,
which I give in Figure 10.

83
Hillocks 2003
84
http://en.wikipedia.org/wiki/Time_flies_like_an_arrow;_fruit_flies_like_a_banana
 Time flies like an arrow; fruit flies like a banana.

Figure 10: The syntax of another pun

It is easy to think of other grammatical entertainments – on newspaper
headlines, for example, which are often grammatically ambiguous as in juvenile court
to try shooting defendant85. A class project might involve collecting examples, and
could then be extended into creating others – an exercise in creativity; and in both
cases the examples would require a grammatical commentary to explain the
ambiguities. A further extension would go beyond newspaper headlines to ordinary
life, with each class member keeping a diary of grammatical ambiguities encountered
in day to day conversation or reading. The main point is to get the class ‘thinking
grammar’ and seeing the relevance of grammar to their own lives.

8 The obstacles
I have painted a picture of a school in which grammar is widely discussed and
fundamental to a lot of activities. As I explained in 3, however improbable this
secnario may seem, it’s certainly not impossible because something like it is already
found in other countries or at other times. The question is not so much whether it’s
possible in principle, but how we might bring it about, given where we are at present.
What, then, are the obstacles? Let’s start with some potential obstacles which
are in fact superable, or even non-existant. One is time for teaching. Teaching
grammar is a relatively new activity, so its introduction is bound to be at the expense
of other activities, especially when the curriculum is already so full. This may be less
of a problem at primary level because quite a lot of grammar teaching is already
taking place. At secondary level it may mean some adjustment and sacrifices in other
areas, but grammar teaching need not claim a large slice of school time if it becomes
more efficient. At present some of the time devoted to it in primary schools seems to
be wasted because secondary schools don’t build on it. In any case, grammar teaching
is an imortant part of the draft National Curriculum, so no doubt school managers and
teachers will find the time needed. I assume, therefore, that finding time is a superable
obstacle.
Attitudes are another potential obstacle. Paradoxically, resistance to grammar
teaching could come from two opposite directions: from those who object to grammar
teaching because they think it has to be prescriptive, and also from those who think
grammar teaching ought to be prescriptive (with a list of ‘common errors’ to be
eliminated) and are disappointed by the absence of this kind of teaching from what I
have proposed (and, more importantly, from the draft National Curriculum).
Personally I don’t believe this is an obstacle at all because prescriptivism is no longer
a major issue either for its supporters or for its opponents. Regarding supporters, I can
mention the recent disbanding of the Queens English Society86; and regarding
opponents, I note that the National Association for the Teaching of English, once

85
http://www.fun-with-words.com/ambiguous_headlines.html
86
http://tinyurl.com/qesdies
famous for its opposition to the teaching of grammar87, has published its own
grammar course88. Clearly I am not suggesting that everybody in Britain is totally
comfortable with non-standard forms, nor that every teacher is totally confident that
grammar can be taught without in any way threatening speakers of non-standard
English. All I am claiming is that most people, including teachers, seem to accept that
non-standard forms are fine in their place, but that Standard forms are needed in some
other places.
Another potential obstacle is the state of knowledge in grammar: maybe we
don’t know enough about grammar to teach it? After all, it is less than a hundred years
since the Newboldt Report89said:
[it is] ... impossible at the present juncture to teach English grammar in the
schools for the simple reason that no-one knows exactly what it is…”
This is another non-obstacle. It’s true that there are serious gaps in our research
knowledge as it relates to school-level grammar and grammar-teaching, and that
we’re short of suitable course material such as textbooks, but there is no shortage of
expert knowledge about grammar, whether specific (the grammar of English, French
and other languages) or general.
That leaves just one obstacle, teachers’ own knowledge of grammar. This is a
real obstacle which cannot be either ignored or avoided. The problem is very simple:
most working teachers, through no fault of their own, came through school at a time
when schools didn’t teach grammar, and those teaching English came through
undergraduate courses which taught equally little grammar, so they don’t know any
grammar except for what they have managed to pick up since the National
Curriculum required it. Interestingly, to judge by anecdotal evidence, a lot of primary
teachers have picked up enough grammar to take children through simple grammar
exercises, and no doubt they are becoming a little more confident every year.
Secondary English teachers are in a different, and more difficult, situation, and
may not have adapted to the new demands as well as primary teachers. One reason is
that their grammar ought to be more sophisticated than primary-school grammar, so
their target is higher. Another reason is the anti-grammar attitude that many of them
were taught during their PGCE year (which I mentioned in the Preface). And a third is
that the need to teach grammar threatens their professional competence, leaving them
feeling de-skilled. I’m guessing, but I do know that’s how I would have felt if I had
been told halfway through my career as a lecturer in linguistics that I needed to teach
a course in literature.
If my guess is right, then some English teachers feel less enthusiastic about
teaching grammar than about the more literary or cultural side of their teaching. It is
for such teachers that I have written this booklet, in the hope of persuading them that
the goal of learning some grammar really is worth the effort and pain, and
(incidentally) that it may even raise their game in discussing literature. However,
before I finish I have some more good news for these grammar-poor English teachers.
Help is available if you want it. There are books90; there are a number of free internet
sites which offer courses in English grammar91; and there are your colleagues in
foreign languages, who may be happy to swap expertise.
87
http://tinyurl.com/antinate
88
http://www.nate.org.uk/index.php?page=11&pub=23
89
Board of Education 1921:289-90
90
Crystal and Barton 1996, Crystal 1988, Crystal 2004, Cameron 2007, Stott and Chapman 2003,
Huddleston and Pullum 2005
91
http://tinyurl.com/ucl-ige, http://tinyurl.com/ucl-ks3, http://tinyurl.com/ucl-app,
http://englishinteractive.net/grammar.html
 References

Aitchison, Jean. 1994. Words in the Mind. An Introduction to the Mental Lexicon.
Second Edition. Oxford: Blackwell

Andrews, Richard 2005. 'Knowledge about the teaching of [sentence] grammar: The
state of play', English Teaching: Practice and Critique 4: 69-76.

Andrews, Richard, Beverton, Sue, Locke, Terry, Low, Graham, Robinson, Alison,
Torgerson, Carole, and Zhu, Die 2004a. The effect of grammar teaching
(syntax) in English on 5 to 16 year olds' accuracy and quality in written
composition. London: EPPI-Centre, Social Science Research Unit, Institute of
Education.

Andrews, Richard, Torgerson, Carole, Beverton, Sue, Freeman, A, Locke, Terry, Low,
Graham, Robinson, Alison, and Zhu, Die 2004b. The effect of grammar
teaching (sentence combining) in English on 5 to 16 year olds' accuracy and
quality in written composition. London: EPPI-Centre, Social Science Research
Unit, Institute of Education - Research Evidence in Education Library.

Anon 2000. Grammar for Writing. London: Department for Education and
Employment

Barton, Geoff and Hudson, Richard 2002. 'Grammar teaching: grind or glamour?',
Times Education Supplement (Curriculum Special) 6-7.

Biber, Douglas, Johansson, Stig, Leech, Geoffrey, Conrad, Susan, and Finegan,
Edward 1999. Longman Grammar of Spoken and Written English. London:
Longman

Blake, Julie and Shortis, Tim 2010. Who's prepared to teach school English? The
degree level qualifications and preparedness of initial teacher trainees in
English. London: Committee for Linguistics in Education

Board of Education 1921. The Teaching of English in England (The Newbolt Report).
London: His Majesty's Stationery Office

Bruner, Jerome 1960. The Process of Education. Cambridge, Mass.: Harvard

University Press

Bryant, Peter, Nunes, Terezinha, and Bindman, M 2004. 'The Relations Between
Children's Linguistic Awareness and Spelling: The Case of the Apostrophe.',
Reading and Writing 12: 253-276.

Cajkler, Wasyl and Hislam, Jane 2002. 'Trainee teachers' grammatical knowledge: the
tension between public expectations and individual competence.', Language
Awareness 11: 161-177.
Cameron, Deborah 2007. The Teacher's Guide to Grammar. Oxford: Oxford
University Press

Carter, Ronald and McCarthy, Michael 2006. Cambridge grammar of English: A

comprehensive guide. Spoken and written English grammar and usage.
Cambridge: Cambridge University Press

Chipere, Ngoni 2003. Understanding Complex Sentences: Native Speaker Variation

in Syntactic Competence. London: Palgrave Macmillan

Committee for Linguistics in Education, Blake, Julie, and Shortis, Tim 2010. 'The
readiness is all. The degree level qualifications and preparedness of initial
teacher trainees in English', English in Education 44: 1-21.

Crystal, David 1988. Rediscover Grammar. Harlow: Pearson

Crystal, David 2004. Making Sense of Grammar. Harlow: Pearson

Crystal, David and Barton, Geoff 1996. Discover Grammar. An enthusiastic and
practical approach to language learning. London: Addison Wesley Longman

Crystal, David and Crystal, Hilary 2000. Words on words. Quotations about language
and languages. London: Penguin

Cutler, Ann, Norris, Dennis, and Sebastián-Gallésa, Núria 2004. 'Phonemic Repertoire
and Similarity within the Vocabulary', in S Kin & M Jin Bae (eds.)
Proceedings of the 8th International Conference on Spoken Language
Processing . Seoul: Sunjin Printing, pp.65-68.

Ellis, Rod 2008. 'Explicit form-focused instruction and second language acquisition',
in Bernard Spolsky & Francis Hult (eds.) The Handbook of Educational
Linguistics. Oxford: Blackwell, pp.437-455.

Gragg, G. 1994. 'Babylonian grammatical texts', in Ronald Asher (ed.) Encyclopedia

of Language and Linguistics. Oxford: Pergamon, pp.296-298.

Halliday, Michael 2007. Language and Education. Continuum

Hillocks, George (2003). reconceptualizing writing curricula: what we know and can
use.Anon.

Huddleston, Rodney and Pullum, Geoffrey 2002. The Cambridge Grammar of the
English Language. Cambridge: Cambridge University Press

Huddleston, Rodney and Pullum, Geoffrey 2005. A student's introduction to English

grammar. Cambridge

Hudson, Richard 1981. 'Some issues on which linguists can agree.', Journal of
Linguistics 17: 333-344.

Hudson, Richard 1992. Teaching Grammar. A guide for the National Curriculum.
Oxford: Blackwell
Hudson, Richard 2009. 'Measuring maturity', in Roger Beard, Debra Myhill, Martin
Nystrand, & Jeni Riley (eds.) SAGE Handbook of Writing Development.
London: Sage, pp.349-362.

Hudson, Richard and Walmsley, John 2005. 'The English Patient: English grammar
and teaching in the twentieth century', Journal of Linguistics 41: 593-622.

Keith, George 1990. 'Language study at Key Stage 3', in Ronald Carter (ed.)
Knowledge about Language and the Curriculum: The LINC Reader. London:
Hodder and Stoughton, pp.69-103.

Keith, George 1999. 'Noticing grammar', in Anon (ed.) Not whether but how. London:
Qualifications and Curriculum Authority, pp.22-32.

Krashen, Stephen 1982. Principles and Practice in Second Language Acquisition.

Levickij, V, Drebet, V, and Kiiko, S 1999. 'Some Quantitative Characteristics of

Polysemy of Verbs, Nouns and Adjectives in the German Language', Journal
of Quantitative Linguistics 6: 172-187.

Levinson, Stephen 2003. Space in Language and Cognition. Explorations in

Cognitive Diversity. Cambridge: Cambridge University Press

Macaulay, Ronald 1975. 'Negative prestige, linguistic insecurity, and linguistic self-
hatred', Lingua 36: 147-161.

Myhill, Debra 1999. 'What do pupils need to know about grammar?', in Anon (ed.)
Not whether but how. London: Qualifications and Curriculum Authority, pp.7-
13.

Myhill, Debra, Lines, Helen, and Watson, Annabel 2010. 'Making meaning with
grammar: a repertoire of possibilities', METAphor 2: 1-10.

Norris, John and Ortega, Lourdes 2000. 'Effectiveness of L2 instruction: A research

synthesis and quantitative meta-analysis.', Language Learning 50: 417-528.

Nunes, Terezinha and Bryant, Peter 2006. Improving Literacy by Teaching

Morphemes. London: Routledge

Owens, Jonathan 1988. The Foundations of Grammar: an Introduction to Mediaeval

Arabic Grammatical Theory. Amsterdam: Benjamins

Perera, Katharine 1984. Children's writing and reading. Analysing classroom

language. Oxford: B. Blackwell in association with A. Deutsch

Pinker, Steven 1994. The Language Instinct. London: Penguin

Quirk, Randolph, Greenbaum, Sidney, Leech, Geoffrey, and Svartvik, Jan 1972. A
Grammar of Contemporary English. London: Longman

Quirk, Randolph, Greenbaum, Sidney, Leech, Geoffrey, and Svartvik, Jan 1985. A
Comprehensive Grammar of the English Language. London: Longman
Reisberg, Daniel 2007. Cognition. Exploring the Science of the Mind. Third media
edition. New York: Norton

Robins, Robert 1967. A Short History of Linguistics. London: Longman

Spada, Nina and Tomita, Yasuyo 2010. 'Interactions between type of instruction
and type of language feature: a meta-analysis.', Language Learning 60: 1-
46.

Stott, Rebecca and Chapman, Peter 2003. Grammar and Writing.

Tallerman, Maggie 2009. 'Phrase structure vs. dependency: the analysis of Welsh
syntactic soft mutation', Journal of Linguistics 45: 167-201.

Trudgill, Peter 1975. Accent, Dialect and the School. London: Arnold

Wason, Peter and Reich, Shuli 1979. 'A Verbal Illusion', The Quarterly Journal of
Experimental Psychology 31: 591-597.

Weaver, Constance 1996. Teaching Grammar in Context. . Portsmouth, NH.:

Heinemann

Williamson, John and Hardman, Frank 1995. 'Time for refilling the bath? A study of
primary student-teachers' grammatical knowledge.', Language and Education
9: 117-134.